Stuff Michael Meeks is doing

Up, off to St Lukes, passable service, good to catch up with lots of the usual suspects. Home, babes to bed, lunch, presents: a number of interesting books.
David arrived, good to see him, caught up with current happenings; good fun. Abandoned by Brothers left / right, watched Michael Palin / bed earlyish.

Up, packed car, woke neighbours, made jump leads from some spare high-current flex, jump started the car ( helpfully the beast has this magic delayed dimming of the internal lights, guarenteed to ensure you don't notice you left the internal light on, and no audible warning of the same ): Toyota.
Drove to Grant & Anne's - stuck in traffic on the M11 for a while - got there, lovely lunch - fine to see all my family, and Bruce, Katy / Ashley, Grant & Anne etc. Wandered the spacious gardens, sheltered from the pelting rain in the (newly mended) greenhouse.
Bite to eat, drove to the parents, stayed up late playing games: 'pictionary' and 'family row' - much fun.

Examined 2nd chimney, a far less helpful shape - considered drilling through the cast-iron facure & re-considered after consulting Father / the wife on the joys of gazing at bolt-heads. Eventually came up with a suitably twisted shape that should stick. Cardboard blank nearly sucked up the chimny by the up-draft. I hate optimising before profiling, but how do you effectively & cheaply heat-loss-profile a room ?
Sandy came around for tea, good to see her, bed early.

Up breakfast, hacked at a new chimney sealing trap-door for the front-room, to save Naomi sleeping in a draft & keep the house warm. Removed much soot from the relevant area; no doubt more will accumulate on the door.
Off to Chris & Cheryl's for a suprise birthday party for Chris (only 30mins late), met Peter Wharton, chatted until late. Home, Gordon sermon while we ate tea, bed to hack while J. bathed. Read the Vista Cost write-up, very interesting / amusing / sad.

Up early, J. lie in. Played with H. & some plasticine hair-cutting thing while killing the last configmgr Address sub-classes, killed memorymodel.hxx finally.
An afternoon of DVD action - first Sindy 'the movie' less lame than expected (suprisingly), babes to bed, then Pirates of the Carribean: The 2nd One - good fun, although, inexplicably the credits showed up 1/2 way through.

Up with the babes, fine breakfast, out for a walk along the Sizewell beach with H. (Sue, Clive & dogs) while the other babes slept.
Back for lunch, coffee, packed, re-arranged the house a little & drove home - roads very clear indeed. Dealt with the snail-mail queue, dinner, bath, bed.

Up early, Father [sic] Christmas generous this year - innumerable pink plasticy things appeared. Out to Church for fine service, more carol friendly one ghastly universalist: Holy is the name I know carol, but older favorites sound.
Back for fine blow-out, rest, then present opening - many interesting books, fine new shoes & so on, great to celebrate His birthday this way: by giving.

Played with babes, over excited wrt. tomorrow, but fun. Sue & Clive arrived for lunch. Out to a crib service with the babes - good fun, albeit too light on carols.

Up late, J. handled the children so I could lie in. Fine cooked breakfast, took Bruce to the hospital nr. Ipswich to have his knee re-dressed. A little hacking in the waiting room.
Back, lunch. Out in the afternoon for a stroll on the beach & to get milk. Popped in to see the RNLI life-boat, very interesting: great to see some of the photos of their work in high seas.
Back, slugging bed.

Up early, mushroom breakfast, to work. Poked at thumbnailing problem, onto crasher on an presentation - chased it down to a cairo clipping issue fixed in 1.0.4, back-ported patch.
On to the performance issue; 3% spent in 'GetBitmapSize' of the same bitmap repeatedly, 10% in BitmapEx::GetAlpha color converting a mask (for some reason). And the other 87% less obviously silly.
Worked late at various curious recent bugs.

Up early, to battle. Mail briefly, killed SegmentAddress, simplified CacheLocation (a struct with a write-only member), starting to look quite nice in places.
Out to Chris & Cheryl's for a drop-in party thing, rather good, albeit horribly cold & foggy. Discovered customer fire-drill on return, prodded various people, played with the relevant files etc.

To work, warning churn reported during the smoketest, unfortunately it doesn't work unless you use <vituperatives> EPM, added some direct make_installer magic to the mix, shame make_installer can only be run from 1 directory, with a very tweaked environment. Hacked a number of nasties out of the smoketest script, and now it installs & runs beautifully, sadly didn't find the reported error though.
Re-worked the cppuhelper patch to please Stefan, did battle with templates again; apparently it's trivial to make simple things more verbose, and far more complicated with them.
Did more gutting on configmgr, partially rationalized the 3 types doing the same thing: TreeAccessor, TreeAddress, and sharable::TreeFragment with some success, more readable.
Dinner, sogged with the Economist for a bit, more hacking in bed - scuppered by switching TreeAddress to a POD pointer type: resulted in several uninitialized members (now there is no default constructor), eventually nailed that.

Up early, lots of paperwork this morning, managed to defeat the multi-hour review that was hanging over me, horay. Thankfully also the cppuhelper re-build completed successfully, tests after lunch.
Lunch, team meeting, poked at cppuhelper crashers. Poked at the MultiTypeInterface...Int32 - turns out though we allocate 392 at startup, we don't actually add any interfaces to them; interesting - tweaked patch for that too.

Up early, walked into town & back; to work. Call with Rene. Poked away at this and that all day - in the breaks got my cppuhelper template evil sorted (thanks Fridrich), and set that building. Phone calls much of the afternoon.

Off to C3 in Cambridge in the morning, a very slick service as expected; messaged via the main screen to help Naomi - inconsolable in creche. Back, bed.
Lunch, watched DVDs with babes in the afternoon, out to the NCC Christmas / carol service & shared meal, went very pleasantly.

Up early, out to the market with stacks of babies, walked the length of the town questing for various presents & things.
Back, babes to sleep briefly, out to Tim & Julie's for a big family bash - great food, good to see all the usual suspects, and great to catch up.
Back late in the evening, bite of tea, bed.

E-mail-work ~all day, some cppuhelper hackerage in the evening to prune the (inefficient) hash_map usage burning a chunk of memory. "STL - ineffiency templatized!".
Interview in the evening with Scott Mace.

JP pointed me at Keith Packard's take on Git which makes me take it seriously for OO.o suddenly, an interesting read.
It seems the mail downtime suffered on wal-3 was a result of some Thunderbird / Groupwise IMAP interaction, and nothing to do with Evolution at all; sorry for doubting you Evolution (& team).
Lunch, poked at misc Gnome bugs. Finally removed the huge & multi-year old internal-nss cut/paste patch, required for back-compatibility for systems without some internal crypto symbol being exported; should shrink the ooo-build archive nicely.
Tried to extract E-trade data for payroll; E-trade is monumentally broken for people without a brokerage account, for some strange reason I feel insecure finding bugs in a web / trading application holding my money.
Banged away at configmgr, ported my re-factoring to 2.1, resolving tons of conflicts from my (now up-stream) string re-work.

Up, poked at mail, out to the pre-school christmas singing fest, rather sweet. Back, fiddled with uber-cheap USB keyboard to get an install: eventually renamed NFS install mount point to avoid hard-to-type keys. Call from NOD - network got broken (somehow), someone fiddling with cabling: bother.
Time to look at the configmgr oddness, found the root cause, some over-enthusiasm in removing (obsolete except for debugging code) accessors, attached patch.
More code-reading in configmgr to attempt to find where the old custom allocator was cleaning up things that were not freed as part of it's operation. Committed configmgr debug redux to CWS configmgrassert, Stefan kindly agreed to QA.
Finally got an install on my small system; it's not just OO.o that is slow here, YaST & co. are not such fun either.
Big 'ole Mexican meal in the evening with Cell group, 13 people or so, up late talking round the fire.

Up early, Bob Sutor's blog is interesting pwrt. the certainty of ECMA approving OpenXML, the real question is what will ISO do ? Florian pointed me at this nice ODF Alliance Fact Sheet, really a set of interesting questions, laying out the arguments here.
Whacked some soda down the bath plug to try to shift the irritating partial blockage; perhaps the fact that there is only 4 inches of run off over a 10 foot pipe is related to the bug. Waited, plunged enthusiastically, no obvious improvement, bother.
One thing that jrb told me last week, that is most interesting is that one of Carl Worth's focuses for Cairo optimisation is OpenOffice.org impress: that's pretty neat, and if so good news for OO.o.
Read through a load of ORBit2 / bonobo patches that Kjartan prodded me about. Lunch, team meeting, good call with Mathias, internal guy with cut/paste problem phoned, hmm.
J. out for a party with the lasses, leaving me with 3 babies to look after.
Our (AMD sponsored) report into optimisation opportunities in OO.o is now available here it is only ~9 metric pages long, with some pretty graphs; the time estimates & rankings are in highly mythical man-days.
Finally wrote up a report from my travel and sent it, grabbed a SL10.2 RC1 to GM deltaiso, and set the apply off - if only I had threaded the right bit of that code it'd be 2x as fast. If someone wants to integrate pbzip2 into the 'deltarpm' tool, that'd make a lot of people happy (some nice graphs on that page too).

Tried to poke at mail, apparently wal-3 is still malfunctioning, so while it worked enough over the weekend to get mail, it's now down and I trashed the cache finding that out, bother. The joys of a non-production POA, another day of pain.
Eventually the mail cache decided to work again, both mhu and rml agree that allocating memory in order to free it under memory pressure is not a clever move with today's OS This does NOT sound useful! etc. On the other hand it used to be useful: mhu pointed out it was great for 16bit windows, and MacOS 8, with simple global memory management and co-operative multi-tasking. But these times are long gone and dead. So, don't hesitate to get rid of this., neat. Unfortunately - other memory savings tend to require more work.

Again GrokLaw banging on Novell: Novell's "Danaergeschenk" once again IMHO confusing tactics with substance, I had a few thoughts.

Open Source, over Open Standards (I prefer the term Free software, but this sound-bite is more chewable). As in the past, I will bang on about Free Software to people, and mention OpenDocument as one peripheral (and good) advantage of OO.o. However a lot of end-users find the Open Standards argument much easier to understand: after all they create data and not code, and since they can't even -read- code (written in this foreign language of C++) they don't value the real Freedom here as much as we hackers do. Of course, many of my colleagues whom I respect at both Sun & IBM take a different view, and that's fine, and they have excellent reasons no doubt. My interest is in re-focusing attention, co-operation and investment on where the -real- value is: the 1000 man years of OO.o rather than the <5 man years of the OpenDocument spec. I want the Free software desktop to succeed - for me if Microsoft adopts OpenDocument this is not a final victory.
Real Openness - I am personally impressed at the genuine openness and neutrality of the OASIS/OpenDocument process, quite apart from its stark contrast to the ECMA process, which IMHO is reasonably termed rubber-stamping ( at least for the 1.0 version, though I'm not optimistic this will change later ). So - this point is at least truthful, OpenDocument really lives up to the 'Open' part.
Multiple Implementations - this is less substantial, by almost any metric there is only one reasonably complete implementation of each standard. That is OpenOffice.org vs. Microsoft Office. Period. Now - of course, in order to claim you are a 'standard' it is often thought necessary to have a 2nd impl. Hence, there are all sorts of "2nd impls" that minimally fulfill this need, these do not bear detailed scrutiny. But coming back to my 1st point, there are a plethora of products, and a diverse vendor community backing OpenOffice.org based products: the fact that StarOffice, Novell OO.o and Lotus Workplace (to name a few) share a common code-base, takes -nothing- away from the real value of standards (for business): multiple-vendors, no lock-in etc. and is an emerging triumph of Free Software. So again, lets talk about Free Software. For those lovers of KOffice as the OpenDocument 2nd impl. I quote (with manful restraint) from the recent LXF88 / Jan07 review of KOffice 1.6 ...(KOffice apps struggled to open MS and OpenDocument files)... [ do read the article ]. And I suspect like many reviews, this was based on a small sample.

Size of Standard - yes a huge & complex standard can be hard to implement, this is completely true. Of course, this is partly true of OpenDocument - it's quite big too, we just happen to have a nice implementation of it already in OpenOffice.org. The general level of mathematical excellence being shown off here is quite incredible though. If you divide >700man years of OO.o code by 700 pages of the ODF spec, you get 1man year/page to implement which seems slightly silly. Also this new SI unit of specification complexity the metric page is rather comical. There are lots of largish specs out there (where no PDF is available, I used print-preview on the HTML in firefox (A4)):

Spec.	Pages
OpenXML - Markup Lang Ref	5219
OpenXML - Primer	472
OpenXML - Fundamentals	173
OpenXML - Packaging Conventions	129
OpenXML - Compat & Extensibility	43
OpenDocument 1.0	706
SVG 1.1	719
XSL 1.1	514
MathML	509
HTML 4.01	389
SMIL	350
OpenFormula	108
XForms	86
XLink	26

So - we can see there is some truth in the huge standard argument here, OpenXML's markup lang ref is larger than the approximate sum of the other standards that OpenDocument references (and of course it's notable that while OpenDocument "makes use of existing standards" (which is great) it's often a small sub-set of them; so a simple sum of the non-OpenXML standards is not a fair comparison. So, perhaps there is a factor of 5x as much 'stuff' here.

Impossible to interoperate with MS - this is the argument whereby we assume the answer and then run around screaming. Of course, it is true that 100% interop is impossible, but OO.o does a pretty good job, and increasingly so.
Screwing Customers - Georg concludes OpenOffice.org should refuse to add OpenXML to its main branch, and we should avoid OpenXML while spreading information about the problems as far as we can.. Of course I beg to differ. The same logic could be re-used for removing the binary file format support. Is that also proposed ? how would that help anyone ? Also - it is emphatically not just Novell working here to improve interop, people with real customers realise this is necessary. Sun of course has been doing a ton of work to re-factor the core, improve interop, re-write filter logic etc. and, as-yet Novell has done almost nothing in comparison, (though of course we're trying to increase our efforts).
Standards - now this is something I really know nothing about. What is and is not a standard ? What is a de-facto standard ? etc. Actually I think (eg.) DVD recording formats are perhaps interesting. Which was first ? which is a true standard ? which has the moral highground ? why there are various combinations of +'s and -'s ?. All I know is - I buy the cheapest, commodity DVD drive I can get, and it (apparently) works with all of them. In the same way I want to see OpenOffice.org become the obvious choice for the best, most inexpensive, interoperable, re-usable, flexible office suite that anyone could want, that also gives you incredible Freedom.

My conclusion is - if you care most about the success and deployment of Free software on the desktop, everywhere; then it is best to be up-front about that, and not get distracted by some transient Open Standards argument. Microsoft can (pretend to?) play Open Standards but they will (or can) not play Free Software. If our competitors switch to Open Standards that might be nice, but when they switch to Free Software - we have truly won.

Lydia over for lunch, good to see her. Call with JP.
Novell needs to hire several OpenOffice hackers, if you think you fit the bill, or know of someone else that does, please encourage them to apply, location shouldn't be a major issue for the best people. I wonder if any Gnumeric / AbiWord (or even KOffice) people are interested.
Quick call with Florian to catch up on his trip - hopefully the photos will show up soon.

Up early, J. lie in, off to NCC - Janice speaking, did well. Out for lunch with Laura & Crayton. Back, put up Christmas while it got dark, nice. Dinner, bathed babes, put them to bed.
Listened to Gordon sermon on The Image of God, fascinating insights into the Garden of Eden, as a type of Temple. Also, amusing to discover all these on-line, having bought the cassette versions of the 1st 4. xBed late.

Up earlyish, off to Suzannah & Clive's - long but smooth drive, sat by the fire, lovely meal, out to see Father Christmas on a rather fine steam train, mulled wine, mince pies etc. Getting into the engine & looking in the fire-box is quite amazing, so big, and so firey etc.
Back to chat with John & Steph & Sam & co. quick tea, and set off home again, with pyjama clad & intermittently sleeping babes. Bed.

Up early, to work - unfortunately the experimental GWise POA (wal4-2) is dead, along with all my E-mail - so, today might be quite productive with no E-mail, will have to use the spam-overwhelmed & un-loved michael@ximian.com address instead.
Tried to do a little analysis on XSLT, of course Google Trends is rather useful, eg. XSLT vs. Eiffel, at least they are in the same ball-game, Ruby kicks it's backside, Haskell is rather comparable, and Cobol is gamely fighting XSLT. So perhaps my semi-rational hatred of it can be justified by at least some market reality.
Finished the report with some help from Kendy, switched to the nice new template and sent it, poked again at the SfxNewHdl issue - seems in fact it implements a hook for std::set_new_handler, and so is (potentially) useful, if in fact it is a good idea. Removed some OS2 & ancient MAC conditionals, cleaned it up, filed the patch, and poked RML. Had a read of the configmgr code to relax.
Dinner good to eat with the family again, and read a bed-time story: "Khyberie: the story of a pony on the Indian frontier" (good stuff). Coffee & web shopping for Christmas, amazed by the Boden website which could almost instantly list the -names- of all houses in the street from the postcode, including ours 'Revel Cottage'.
Found Gordon's rather nice Historicity of the Bible slideshow, while digging for books to buy.

4 adjacent seats to sleep on on the flight back, better than 1st class. Some work on the train, picked up by the darling wife & daughters. Home, slept like one dead, woken by door-bell, J. lost her key, glad I'm home.
Lunch, Cheryl came over, chatted for a while - back to work. The unusable kernel with I/O occuring bug rumbles on inconclusively.
Chewed through the memprof output for my report / analysis. Hit a wonder of a class 'SfxNewHdl' that (apparently) allocates 512 Kb and holds on to it, just in case there is some out of memory condition, in which case it can free it - to release some memory to throw up a warning dialog. A cunning plan indeed, unfortunately a quick glance suggests it is never used. Looking forward to actually doing the hacking to fix some of this.
Worked late, finally drew all the bits together into one document, looking nice, bed 11pm.

Up at 5am, poked web / mail. Fixed some localtime to gmt based accounting in my evily hacked perl HTML -> RSS conversion script, hopefully now it advertises the right day of the week for a given date.
Petr did a great job packaging the OdfConverter, now available for SLED10 at http://red-carpet.go-oo.org; test docs. This is of course a stop-gap measure, the goal is not a separate converter (using XSLT) but high quality built-in OpenXML support, re-using & re-factoring the existing Microsoft binary filter code. Of course Mathias Bauer points out the Sun writer team is doing some great work re-writing the existing code here to be more flexible & elegant - read Henning's nice post about the design here.
Decided wearing the FUD Fighters T-shirt Simon Phipps posted me couldn't be more appropriate today.
Team meeting - an appallingly bad and expensive hotel line. Met up with Kay afterwards, chewed over various current issues in the code, lunch, and the afternoon. Read a nice analysis paper from Noel. Interesting talk with Leon in the afternoon.
Caught up all too quickly with the Teichman, Berkman, Garret, Gordon etc. good lads. Synched with Miguel sadly in rather a hurry, and off to the airport (of course flight delayed).
Poked at Ricardo's latest yast-gtk work, looking really clean & nice.

Up early, meeting all day - initially constructive, went down hill, lunch, went up-hill. Some great guys though, good to see JRB again.
Out for dinner in the evening at the top of the Prudential tower, lovely view, good food etc.

Up early, packed frantically, synched critical mail etc. lovely wife dropped me to the train. Chewed the ignored mail backlog on the train.
Call with Kai Backan, interesting, call with (cousin) Andrew, apparently recovering nicely from his climing accident.
Virgin Atlantic - typical nanny-airline, insists on showing you a ton of stuff you don't want to see before turning on the inflight entertainment 30mins into the flight. Of course, as normal the system crashed shortly into the 1st movie, necessitating a total system reset (another 45mins) - yet more embarassingly one of the odd diagnostic screens shown to all customers contained the simple legend Launching ./airsurf, sigh.
Got on with writing a nice report on the OO.o resource footprint - good to have the chance to do that all in one place.
Landed; to the office - met up with Miguel, once again amazed by PJ's charged polemic Novell "Forking" OpenOffice.org - something that we are emphatically not doing. The 'Patent Office' jibe is just totally inaccurate wrt. the MS Open Spec. Promise - which (AFAICS) was inspired by Sun's equivalent for OpenDocument (more power to Sun). All the work is public wrt. using the translator, in fact we do no closed-source work on OO.o. On the other hand, I personally am looking forward to working on rather better built-in OO.o OpenXML filters, re-using the existing nice filter logic for the binary file formats: .doc etc.
Out for a fine dinner with interesting company.

Up, off to Rock Baptist in the morning, followed by lunch with James and Kate Williams. Turns out James has one of these newfangled blog thins noearthlycity. Also, a link to his recent Cambridge Paper on Humor, Scripture and Christian Discourse. Great meal & time to catch up.
Back, cheese on toast for dinner, DVDs for babes in the evening while knocking out some party invites for Naomi's birthday.

Up early, off to the market with a full complement of babes; stocked up on veg etc. Back, some cleaning up, Barbara & Colin arrived for lunch.
Iris came over, helped re-arrange her flights while watching the Prince of Egypt variously. Talked to Barbara & Colin by the fire for the rest of the evening, with occasional programming assistance for Thomas. Bed early.

Worked on details for Boston trip next week. Some more profiling of various bits of performance / memory usage. It seems configmgr2, as expected is still the top of the profile after linking, must get around to finishing my re-factor there to make the code readable.
Bed early, still exhausted generally.

Up early, talked to Tor - some problems with the mono staticizer - it appears that though you staticize it still needs a 'machine.conf' file where it expects it. Slowly calmed the mail queue down from it's excitable state.
Wrote some slide notes for Alan Clark (volunteer hero of the week). Lunch. Plugged away at various things. New interesting hardware arrived to play with.
Off early for dinner, back - more misc. IM, started doing a prolongued install fest on the new hardware. Hacked relocstat to dump method relocations, assuming that all method & thunk relocations are vtable related we waste up to ~9Mb with relocations we shouldn't really need.
Got some nice cachegrind output for OO.o startup, really need to have it on SL10.2 though.

Up in the night, knocked off some mail, back to bed. Up later - interested to see this interesting 'strip mining' analogy deployed, in reaction to this sort of thing.
Irritatingly my 64bit build machine / file-server broke-down overnight, next time, I'll buy Dell instead of supporting the local Mom & Pop shop. Set off some RC1 DVD download goodness.
Call with Simon, poked mono list to find Environment.ExitCode - neato, added it to the OdfConverter so we can detect failure nicely.
Lunch, interesting call with Jurgen, Andreas, Stefan & Noel, call with Alan Clark. Poked despondantly at dead hardware, bother.
Started on some fun performance work, measuring things is always fun.

Up lateish, dropped Mother & babes at pre-school, to work - dug at mail variously. , various calls. Lunch, team meeting, more calls.
Wife arrived home, fed baby (now apparently better) and went to bed (ill herselt). Did a scad of OMT peer review at lightning speed, knocked off for dinner.

Up early, dealt with babes, hacked in the kitchen while colouring goes on. Today is the "Objectives Management Tool" (OMT) deadline. In the same way that an organisation couldn't possibly know what it's purpose was without a mission statement, employees cannot work effectively without a formal objectives management process (or something).
Put babes to bed, hung washing out, put nappies on, back to the OMT. The OMT amazingly also doubles up as an employee patience training device, with added helpful anger-management experience features. Not content with being slow & hopeless, it likes to wait until you've entered a load of text before some random key combination encourages the browser to loose it all, stretching stuff. Isn't the web 'paradigm' so powerful wrt. entering tons of text into text areas ? Eventually got the nauseating self-praise entered.
Merged, tested & re-filed Radek's 64bit converter fix. Woke babes, lunch, off to Addenbrooks.
Talked to J. for a while - hasn't been sleeping much poor dear; looked after the children for a while while she slept - Miriam looking better, and less feverish thankfully, J. getting a temperature though.
Back late, fish & chips, babes to bed, stories etc. call with Mum & Dad, offered to come and help: kind.
Back to work. Poked mail, poked Guy, hassled JP, poked Miguel / Mono guys on the evils of LLP64 Win64 vs. LP64 Linux, what fun. Toshok kindly pointed me at FYIReporting as a nice sample cross-platform SWF / Mono app to demo, neat. Eventually decided it's better to actually work instead of hassling people.
Isolated (thanks to Peter Vandenabeele's persistence) the OdfConverter bug - to somewhere in the ParagraphPostProcessor, and fixed it - another trivial case of the WriteFullEndElement method not getting called in Microsoft land (for whatever reason), prolly some deeper evil lurketh here. Up-loaded odf-converter-0.0.4.
Mother arrived late.

Up at 1am, Miriam unhappy, fever recurred, J. unhappy. All headed off to hospital, back to sleep. Up early, J. kept in, Miriam on intravenous antibiotics - unclear what the problem is, no obvious cause.
Got babes dressed, tried to get the ham stewed in apple-juice etc. made. Lift with Dave & Cathy to church, did creche - back with Cathy, enlisted guests in various washing up / table setting duties. 15 for lunch - hard without J.
Cleared up, Mel volunteered to lead the prayer meeting this evening instead of J. Nick gave me a lift to Hospital while Jackie baby-sat, nice. Good to see very tired wife, and baby, fever apparently subsided after a day of it, reluctantly left them there. Back, bed late, tired.

Up, a day of construction - got the medicine cabinet finally up on the wall - horay. Peeled back the bathroom carpet to investigate the dodgy sounding floor-boards.
Out to Daniel's birthday party - interesting people, lovely food for the babes, fun & sticking. Back, babes to bed.
Discovered that mostly the bathroom floor-boards are held up by the copper piping underneath (urgh), routed channels out in the boards to accomodate the piping, attached supports variously to (horribly inaccessible) joists to support them etc.
Slugged in the evening, bed.

Up early, call with Florian & some of the Writer guys, very constructive. Poked mail, more mail etc.
J. off to hospital with unhappy sounding Miriam, left holding the fort. Call with Nat, burned time with IM, call with JP.
Managed eventually to hack some more pleasant exception reporting into Mono's XSLT runtime, now I know where the problem is but I still don't understand or like XSLT: time to learn something new.
Ian Pratt pointed me at the interesting DiskSim impl. for modelling disk performance, nice man. Hopefully with some nice condensed application traces and so on, it'll be possible to model the Linux I/O stack & prototype changes to various algorithms.

Up late, poked at mail, trawled through the mail progress of Kjartan sorting out a ton of old/stale/fixed ORBit2 bug reports. Chewed couple of days of mail. Call with Noel.
Setup various conf calls, talked with JP.
Stefan Bergmann pointed me at his (apparently) canonical lets fix rtl:: String classes mail. In the 3 years since then, we seem to have made little progress at making rtl::OUString as useable as the tools string classes, so far we have 3 new methods:
```
    explicit OUString( sal_Unicode value ) SAL_THROW(())
    static inline OUString const & unacquired( rtl_uString * const * ppHandle )
    inline bool endsWithIgnoreAsciiCaseAsciiL(
        char const * asciiStr, sal_Int32 asciiStrLength) const
```
Clearly we need to move a little faster to make the rtl String classes more useful if we mean to deprecate the tools equivalent.

Up early, off to the Museum, met the various worthies here, prepped & did a SLED10 demo. Hacked at the OdfConverter while listening to a 'Why Groupwise is best' talk. Disappeared into an XSLT morass, urgh, urgh, urgh.
More hours of grappling with Mono: the lack of helpful information on exception with a full XSLT processing stack-trace dump is particularly painful. Perhaps few people try to do so many 'unusual' things as in this converter. Dogged by un-debuggable mess. Switched to some more productive configmgr2 re-work work.
Home, dinner, cell group, bed late.

Up early, Noel pointed me at this piece of what is (technically) known as the purest crap. Anyone who is at all confused can look at the relevant FAQ (Q4.) which I quote in full here:

Q4. With this agreement, will Novell include Microsoft patented code in its contributions to the open source community?

No. Novell will not change its development practices as a result of this agreement. It has always been our policy in all development, open source and proprietary, to stay away from code that infringes another's patents, and we will continue to develop software using these standard practices. If any of our code is found to infringe someone else's patents, we will try to find prior technology to invalidate the patents, rework the code to design around the infringement, or as a last resort remove the functionality.

Novell is committed to protecting, preserving and promoting freedom for free and open source software.

This is the basis that my team work on. The VBA compatibility functionality is just an extension of existing MS Office interoperability work. I suspect that the poster is a pre-gratuity equine dental examiner looking for custom. I mean, even the premise that VBA (a simple Basic) is in some way revolutionary, innovative or novel is rather a stretch. The nice piece this was based on, and is rather more interesting is OpenOffice.org Calc adds support for Excel VBA. Of course, if you like Deja-Vu you might want to read this after the Sun announcement in which the heebie-jeebies are going at them instead.
Another meme I've been reading recently in the Argh, everything is a trap ! vein is the myth of Microsoft infallibity: the argument goes something like this:
- Microsoft has never made a mistake
- => anyone making a deal with them will get shafted
- => we're all going to die
- => run around screaming.
While, for sure Microsoft has made some astute business decisions in the past, it has also been lucky in some cases - and also rather stupid in others. Of course, taking the infallibility argument to it's logical conclusion we should not even bother trying to compete with Microsoft, but all roll over & play dead. Anyhow - luckily I beg to differ, and present two counter examples, that I happen to remember:
- Sun & Java - Microsoft licensed Java, news.com writes:
  
  The germ of the suit began when Microsoft took out a Java license in 1996. Sun contended that Microsoft quickly began to run afoul of the licensing terms and filed the initial lawsuit in October 1997.
  
  So here we see Microsoft making a really stupid move, signing an agreement they couldn't keep, and probably didn't need, and what was the end game ? they got to buy themselves out of it for $20million, got a load of injunctions slapped on them, had to retro-fix their product etc.
- Lindows vs. Microsoft - the Wikipedia article is quite illuminating:
  
  As part of this licensing settlement, Microsoft paid an estimated $24 million cash (for a case that Microsoft itself brought), and Lindows transferred the Lindows trademark to Microsoft and changed their name to Linspire.
  
  And now of course, we all know the Windows & Office trademarks are indefensible, one rumour I heard had it that the settlement itself was intended to be secret, but SEC regulations mandated disclosure & this was missed by the MS lawyers.
Unfortunately the myth of MS infallibility is rather useful to them in a market where leadership is critical, and can move the market.
Got to the mail queue, submitted another patch to CleverAge, seemingly they merged my last lot of changes, nice. Lunch, IM chat overlapped with the team meeting (sadly).
Rushed to train station, call with Florian. Worked on the train, got to 'Walkabout' - an interesting hotel, but cheap - dinner, reading The Hiding Place (Corrie Ten-boom), more demo prep in the evening.

Up early, noticed I had removed my system openoffice in a fit of carelessness, fixed that. Got Ken Muir's nice Netware/Groupwise POA snapshot installed and even working, along with his new updated alpha Java client; prepped for Groupware Users Group conference / SLED10 demos on Wed.
The YaST team very kindly lent us Rene Engelhard for some applied OO.o hacking for the next 2 months; great, lots of things for him to do here I think, thank you Edith. Hacked away at talk, call with JP.
Dinner, finally got to updating & fixing up various converter conflicts post update. Played with mkbundle, managed with some linker hackery to get everything bundled into the one ELF binary, including the ugly libzlibwapi mess.

Up lateish, off to St Lukes - lovely to see so many old friends and catch up a little. Back for a lovely roast meal with the parents, while the babes slept. Fine afternoon, watched part of Babar on DVD, installed OpenOffice on parents PC.
Smooth drive home in the evening, packed children into bed - Naomi seemingly much happier with her teeth, cold, ear-infection tripple whammy.

Up early, breakfast - off out to the park with the babes, lots of fun there on various climbing frames etc. Back - put the babes to bed. Checked mail.
Lunch with the parents (returned from a talk), had a fine afternoon slugging, admiring the handiword, fine cake for afternoon tea etc. Didn't sleep so well.

Up a good chunk of the night / slept with a little Naomi with very short painful breaths, teething and with an ear infection - very sad. Off to hospital at 5am, not much sleep.
Awoken from numbness - checked vtune builds, better detail from instrumenting a few libraries, but the tool unfortunately sucks for usability compared to 'speedprof' (or sysprof, or kcachegrind etc.); didn't manage to find good call traces of where the phenomenal Win32 slowness comes from - however, hopefully can be extracted from cachegrind on Unix by simple call count accounting.
Set off a huge download over the VPN, packed the house & girls into the car, set off for the sunny south & my parents. A little hacking on the road.
Conf-call with some great Sunnies, call with JP / Florian, digging at documents. Got my 1st successful install-set from the BuildBot [ sadly not from the canonical SuSE-7.3+stuff ABI base ], but a milestone in itself; neat. More phone dither. Call with MikeLeib. Dinner, more mail during dinner.
More work in the evening. Knocked up a dist: rule for the odf converter, now only 170k small.

Poked mail. Did various queued tasks, re-submitted XMS expenses etc. Re-started the 'tag-list update' EIS -> text file script that (seemingly) jammed earlier this morning. Perhaps should add a cron job to kill it & remove the tmp files (sigh).
Manually hacked the 'tag-list' to add my OOE680 CWS, unpleasant in the extreme, banged the BuildBot again (apparently tag-list doesn't do the job for non-SRC680 masters). Prodded column again.
Dinner - left with the babes to sit on while J. out to bible study; Myriam letting rip. Unwound an interesting threading / a11y issue hurting Ekiga for Padraig. Sent status report. Set win32 builds running

Up, read LXF for a while - enthusiastic interview with Nat, nice. Poked at mail. Jdub tells me that the ClassPath exception is an LGPL-like workaround for the "complete object files" blurb in the LGPL built on the GPL; sounds reasonable - but having some harder core description of this exception somewhere would be most helpful.
Dug at Pavel's bug, with no success, can't provoke it in my build, not clear what is going on, downloaded the Sun build. Finally discovered the solution for the magic 'Quickstarter enabled by default' cases: the 1st start wizard turns the quick-starter on (magic).
Good calls with Kelli, then JP. Out to cell group at Nicki's, got a little carried away by the Bible study & over-ran a bit. Bed late.

Up early, to work - plugged away at mail, then various mono pieces; collected H. from creche. Set off a build of the 2.1 branch to find the bug Pavel reported.
Team meeting, spammed people with action items. Poked at mkbundle, with lots of patient help from Paolo & Miguel discovered that setting MONO_PATH is the cause of all my problems, unsetting that, using mkbundle2, using --deps (and amusingly re-setting MONO_PATH for that command) yield a beautiful standalone elf binary (well it (sadly) requires one extra libzlibwapi.so) that converts OOX to ODF and back.
Kay - revisiting the OO.o Java policy looks sensible to me - and not to look a gift horse in the mouth, but I'm confused: Java is GPL not LGPL - and there is a world of difference and pain here. Of course there is the ClassPath Exception, but what does that mean for OO.o ? is this essentially a 'Lesser' (LGPL like license ?) if so, why not just use the LGPL ? does it mean 3rd parties will no longer be able to write plugins ? is there a good explanation of the ramifications of the ClassPath exception ? will Sun be distributing it's own Java under the terms of the GPL ? the ClassPath licensing exception page is really rather vague. On the other hand, if this gets unwound, it's great news that perhaps we can make OO.o more buildable and usable, and not be so concerned about Java code drifting into OO.o.

Wrote LXF column prototype, dispatched team meeting agenda, more mail, bits of paperwork. Lunch. Lengthy call with JP. COBE flash training things, hard to remember if the quiz questions are the same as last time.
Buffed up the OdfConverter Makefiles a little, found / fixed the Mono System.Xml bug hurting the nice trick of packing the relevant XSLT files into the binary itself.
Dinner, bed early - up in the night - installed VTune, poked at mail. Ran VTune on OO.o/Win32 - it seems to show a great whack of configmgr2 burn, but the results seem rather difficult to interpret unfortunately.

Up earlyish, fed babes while J. slept, off to NCC Simon Matthews speaking - on running the race; went through a (couple of) passages and explained them well.
Back for a fine lunch, more applied inactivity, and catching up. David left mid-afternoon, watched DVD and read to Naomi while J. & co. out at junior Corf-ball training. New strategy with Myriam - resulted in getting to bed earlier, nice - read lots more of In Search of Stupidity, an amusing marketing perspective on the tech. industry.

Up early, looked after babes while J. went to the market, tidied the house. Out to Chris & Carol's for a 'Chocolate Party' for Naomi's 4th birthday. Good to see Chris' setup and their fascinating house, fine chocolate too.
Home, David arrived - really good to see him, slugged by the fire much of the afternoon and chewed life over.

Up lateish, to work - hacked away at the OdfConverter - with Florian's XSLT skills working on my brain-damaged debugging code, and the broken document named "valid.docx" - a miracle we achieved anything. Eventually though we got the first .docx converted to .odt and loaded in OO.o: nice.
Took Florian to the station, and then on to Cambridge (cancelled train), back - call with JP. Finally replaced my Icelandic keyboard for a UK keyboard, 6 months to ship the right keyboard to me ? impressive work from Calyx.

Up early, a day of talking, trying to understand the writer layout design, and various other bits.
Out to Cambridge in the evening for dinner at CB2 with Julian Seward.

Up early, a new set of helpful FAQs on the MS/Novell deal as it relates to open-source arrived: good. Also a more official looking Novell statement on our OpenDocument involvement arrived.
Poked at another proposed e-d-s change, of course all changes to remove (unused) API are good, as long as the major .so version isn't changed - breaking the OO.o integration. Massaged bugs, call with Kai. Really it seems software engineering is by far the easiest piece of the job; fixing or working around organizational dysfunction is by far the more difficult & (perhaps) interesting piece.
Out to NOD to fix their networking, back - call with Jody.

Up early, dug at mail, considered life.
It seems some people got scared (for inscrutable reasons) about Novell's commitment to OpenDocument, OpenOffice etc. - just to re-affirm, I'm a huge fan of OpenOffice period. There are also some really nice things about OpenDocument, and we'll of course continue to invest in improving it in the areas where it's deficient. Hopefully that process will accelerate with both Jody and Florian on-board, and we'll end up with an even better OpenDocument standard.
Separately, I was somewhat amazed to see Simon Phipp's comments on the Novell/MS announcement. For a while I have been hearing (from Sun) how great StarOffice is IPR-wise due to Sun's settlement (cross-license ?) with Microsoft. Of course, at some level Simon is a lone-voice of sanity inside Sun, but it seems strange to be advertising on on hand the joys of a proprietary product (StarOffice: with IPR protection) and simultaneously slating Novell's moves to protect customers using our Free Software product. Miguel has a nice write-up of various other typical coments.
Poked bugzilla, internal admin, looked at an a11y leak with Padraig, poked at Unit tests again. Team meeting, posted action items. Phone call with Boston.

Up, poked mail, faxed paperwork, poked XMS etc. Off to Stanstead to pick up Florian, back lunch. Call with John McCreesh, An hour and a half chat with JP. lots of time with Florian, no hacking (bother). Dinner, bed.

Up early, fed babes, off to NCC - quite a good showing, talk went well (thank God), last of the 10 Commandments.
Back, lunch, out to see Lenny at Addenbrooks - poor dear, suffering rather, read with her a little & prayed. Back - slugged with kidlets, watched (Mole) DVDs, etc. bed early.

Up early, off to the market with the babes, vegged up, home, met Bert on the way - on a walk for exercise to his wife's grave (so sad). Grandparent's arrived - played with babes, hung washing out, etc. while J. & Mum went out to a craft exhibition in town.
Slept by the fire some of the afternoon, fire-works in the evening with the Grand-parent - of the non-bangy variety: lots of showers of sparks etc. Stayed up late writing talk on Coveting for tomorrow.

Up early, call with Florian. I forgot to mention yesterday (amid the bustle of interesting happenings) that he has joined Novell. Florian (as CTO OpenDocument Foundation) has lots of experience we want to bring to bear on our OpenOffice.org development and interoperability efforts, particularly in the writer core. Florian of course has contributed to OpenDocument as a member of the TC, and of course I envisage that role continuing, though no doubt he will also make a valuable contribution to the OpenXML / ECMA effort. Either way, it's great to have Florian on board, and back working to improve OpenOffice.org. He'll be blogging soon.
Woot. Petr tracked down the evil Python crasher plaguing OO.o component registration. Hacked away at some Mono / C# / XSLT code. Re-printed & faxed updated expense (not a good use of time) - all for some trivial adjustment, sigh.
More hackery - got the OdfConverter at least compiled and run-time crashing; good. Vaguely amused by the various predicitions of doom and legal nightmares. RIP Nuclear Power - killed by the heebie-jeebies.

Up early, N. very upset after her 'duch' was confiscated in an attempt to stop the process of her climbing up the ladder into her sister's bed (waking her) and getting stuck up there, remonstrated, returned to sleep.
Up early, poked mail, started to see some hacking at the end of the tunnel - nice. Lunch. Call. Chewed over the optional arg patch with Eike, need some more tests. Generated some warm-start numbers for Malte of OO.o:

WinXP 2.0.4 Lin 2.0.2 (Nov) Lin 2.0.4 (Sun) Lin 2.0.4 (Nov)

Avg. Time (secs) 4.0 2.4 1.69 1.55

Clearly we're getting faster, at least on Linux - and clearly something really dumb is happening on Windows, I suspect configmgr's locking fetish personally; hopefully VTune will show us something.
Started doing some hacking with Mono - soon fell foul of the jumbled impl./interface concept, #mono had some helpful suggestions, a simple grep 'public' being great.
The Announcement ( MS) happened. It's been a privilege to be involved with a small part of this process, and as always working with lawyers (those word-hackers) helps sharpen the mind. It's great to see some of the full scope here.
Inevitably people will have some really good questions here, and to save my fingers I thought I'd point out a few points.
- Why help Microsoft with OpenXML interop. ? OpenXML sucks, OpenDocument Rocks !?. So several thoughts:
  - This should not be a surprise - Jody Goldberg (on my team) has been working hard for months with Microsoft and others on the ECMA process. At one stage there around 1/2 the open 'issues' wrt. improving disclosure (and hence the spec.) came from Jody. I for one am proud of the job that he did there, an (ongoing) investment that will yield better interoperability for years to come.
  - As I have said for many months now, focusing on an 'Open-Standard' of ~700 pages written by a small team over a short period, is to miss the staggering value that is found in Free software. OpenOffice (as anyone who tried to start it recently knows) contains millions of lines of code, and a staggering investment of thousands of man years of sweat, tears (and perhaps blood). It's localized to umpteen languages, has deep help, scripting, accessibility, interoperability; it's just an immensely feature rich and powerful product.
  - To re-emphasise this, the value in OpenOffice.org is not what file format it supports (eg. we want to add good Lotus Word Pro support) but that it is truly Free software, that gives people critical Freedoms. An open format is anyhow implicit in the native file format of any openly developed Free software project.
  - Telling people about open standards, instead of Free Software is easy - 'normal' people generate data, not software so they understand, but it sells them radically short. In my view better interoperability (with any and all formats) strengthens Free Software, quite without the obvious pragmatic benefits to users & customers.
- Why do business with these scum ?
  - It's true there is a widespread perception of unfair business practice from Microsoft out there, but my experience of working in the ECMA process with the developers, has been of meeting a (to my mind) mis-directed, but equally passionate world-view based around the love of their technology.
  - Broadly, I think it's fair to say there is a certain kind of person that loves to solve complex, technical problems, and I like that kind of person. It's also interesting to note that the average Microsoft (from my small sample) political viewpoint is -way- to the left of the average Novell Free software developer (perhaps a statistical aberration but ...). So, in a nutshell, they're good guys, if mis-directed. The great news is that we can help change that direction and get these guys addicted to the Free Software model.
  - One couple it was fun to meet, both on the Office team, obviously in love, confided in me that they had delayed their marriage to meet the Office 12 schedule: is that dedication ? Let's hope Wedding 2007 will ship on time; but imagine if we can help focus these guys on improving Linux <-> Windows interoperability, and in time Free software for it's own sake.
- What does it mean for OpenOffice ? - my hope is over the long haul: better interop, more bodies hacking on OO.o, wider penetration of (Novell's) OpenOffice into the enterprise, and more individuals able to boldly hack on Free software.
- What does it mean for Hackers ? - of course, I'm pleased that our team got such a great formal IPR covenant for individual developers from Microsoft. For sceptics that think this is a pure gesture, it's always surprising to me how a few key people seem to pop up again and again in Free software, and not everyone has the 7 year stamina that can be required, the RIAA demonstrates the danger well.
- What does it mean for Novell ? - I'm pleased that it seems Microsoft will be distributing lots of SLES coupons, the more the merrier. Of course Nat and Miguel who helped setup the deal have a clearer view.
- What is this Translator ? - it's the early stages of a open-source project to make a standalone bi-directional Open XML to OpenDocument converter. See SourceForge: odf-converter. What is important to me is not the set of design choices here (a standalone XML to XML converter: though that may be useful for other Free software projects such as Beagle, or it's capabilities: a sub-set of Word only so far). What is important is the end-goal of getting substantially better MS Office interop. (with OpenXML) into OO.o.

	WinXP 2.0.4	Lin 2.0.2 (Nov)	Lin 2.0.4 (Sun)	Lin 2.0.4 (Nov)
Avg. Time (secs)	4.0	2.4	1.69	1.55

Up early, spent an hour trying to unwind some appalling internal mis-communication, egad. Generated a patch for Tor to help with some more QA. More mail. Posted interview questions, decks starting to clear.
Started doing a round-up / analysis of the spec. thread - seemingly there is a lot of common ground, and opportunity to improve. Off to the Doctor's - amazing, you arrive in good time for an appointment, 15 minutes of waiting later, no show.
Cell group, call in the middle, as always lots to pray for in our little group.

Up early, poked mail - interesting; clear communication is extremely difficult it seems. Team meeting, lots of good things ongoing as always. Lunch.
Encouraged by the spec. discussion on the OOo dev list, it looks as if real communication is starting to happen. Late meeting, bed.

Children up early, forgot to change their light / timer back the hour (doh). Poked at mail, knocked up a quick script to run/quit OO.o automatically. Poked at the gtkquickstart2 CWS.
Sold our Vauxhall Corsa after lunch, good to get it liquidated and the mortgage down a little further.
More spec. process pokeage, cloph caught me being particularly dim with my scp2, thanks. Filed expenses frantically, managing to improve the muscle-memory and get to grips with XMS's Hotel itemization, nearly in time for the next system to arrive.

Off to NCC a full hour early, it seems the clocks went back (bother). A family service - rather good this time, better seating, several visitors, fun.
Back for lunch, slept afterwards, up - took H. & N. out to watch the beautiful sunset while having a picnic. Wonderful orange / pink / red colors (much beloved of little Girls). Nice to see the moon, but too much cloud for stars.
Bed delayed by much screaming of small people.

Up early, out to the market with a full compliment of babes; bought several lbs of good things, and back. Dithered about screwing a rather hefty (set) of medicine cabinets to the (somewhat flimsy) bathroom wall [ half brick, and part thin cinder-block on timber.
Left the children playing downstairs while investigating the finances, eventually went down to discover the cooker with all rings glowing, oven on and 2 girls shrieking with delight as they loaded various toys into it. Spankings all around, confiscation of precious toys, much heated explanation, repentance etc. It -looked- like Hannah was conding if not egging on what Naomi was doing, hence the "You're older & should have known better" adage familiar to older siblings.
Babes Mole DVDs in the evening, hacked a little at yast-gtk while that was going on, can now produce ISOs that boot and at least render some stuff, screenshots to follow. Reported Novell Forge downtime, fixage deferred until Monday (urgh).
Watched Pirated of the Carribean: The Black Perl in the evening, emminently watchable.

Up early, call with Alex (AvengerMojo), poked mail, more mail, lunch.
Read the specification thread, tried to be succinct, broadly failed, read a specification (almost a 1st) noticed it riddled with errors etc. neato - yet another validation of the concept of (code) review improving quality more than the generation of yet more (buggy) content (?).
Lunch, back to finish off.

Up late, to the conference - wander the floor meeting new & interesting people. Good to see Eddie Bleasdale again. Lunch with Zaheeda, set off home.
Hacked yast2-gtk on the train, splitting the wizard help out into a separate dialog with search is quite a nice touch I think.
Call with the mother-ship in the evening, take-away pizza with the sweetheart, and (unusually) out to the local Pub in the evening with David, bed late.

Up at 6:30 - off to Kensington Olympia, LWE-UK time. Tested demos, (all except for the trusty camera), did a nice desktop demo / talk / thing; went fairly well.
Out for pizza with Owen & Adam. Interviews & bits in the afternoon. Good to see Alasdair again, and chew things over with Ian Pratt: a man with an amazing knowledge of the workings of disk drives, and some scary tales to tell.
Hung out with the Novell lads a little, to hotel; checked mail, responded to this / that, oddly smart looking dinner jacket on, off to the UK Linux & Open Source Awards, good to see Chris DiBona. Sat with Thomas Wood & his fine lady, interesting evening. To bed late.

Up early, mail, lengthy call in the morning, team meeting. I wonder how global the aphorism Never look a gift-horse in the mouth is.
Started to prepare for tomorrow's LWE talks & meetings rather late in the day, pleased generally with demo prep - far fewer bugs to fall over than before: clearly not demoing the cutting-edge this time. Bed late, up in night with N. being naughty.

Up early, ORBit2 bug mail triage, Kjartan fixing things / triaging long-term bugs left & right again.
Read a little on Quality - it's been a while since I had to look at some of the formal approaches here; [ health warning: WikiPedia is dangerous for intellectual's productivity ]. Demming's 14 points are interesting. Particularly amused by point 11. Eliminate numerical goals, numerical quotas and management by objectives. Substitute leadership.. Contrast that with "Gibb's Law" (Peopleware) Anything you need to quantify can be measured in some way that is superior to not measuring it at all Conclusion: the Quality space is riddled with local maxima, punditry and very different (perhaps equally good) approaches. Consequently pragmatics has to prevail, as it does at this fish wholesaler
Moved the mail mountain a little. Call with Kai, back from holiday. Picked up the car from the mechanics - seems better in several ways. Did some interesting (for once) paper-work. Finally got round to up-loading the dmesg output for the I/O badness I see.
Poked at some bugs, split out the 3rd bug from the quick 2nd starter issue - a number of places in the OO.o code inadvertently use the (unhelpful) UniString constructor:
```
String(RTL_CONSTASCII_FOO_LONG_MACRO_PARAM("foo"))
```
which due to the miracle of polymorphism this ends up doing:
```
UniString( const sal_Char* pByteStr,
           rtl_TextEncoding eTextEncoding,
           sal_uInt32 nCvtFlags = BYTESTRING_TO_UNISTRING_CVTFLAGS );
```
with eTextEncoding == 3 == RTL_TEXTENCODING_IBM_437 (or somesuch); not at all ideal.

Up early, to NCC - did Creche; God loves Children tried to convey it to the under 3's - eventually we mostly agreed that they are in fact children, not sure much else sank in.
Home for lunch - 14 people around the table + 2 babies. Good fun, talked late into the afternoon, poked at Chris's SDQS website. Various complaints from children all over the place, bed late.

Up at 3am to tidy up some loose ends from the day. Also got my 1st negative comment on my photos, so far fairly few spontaneous private comments (from outside Novell) a complete summary (excerpting & anonymizing):
- Community hacker - thanks a lot for your las blog entry ... i see i'm not alone
- Code Contributor - I had to just drop by and let you know your recent blog entries had me ROTFLMAO. :_D ... You have to laugh, or else I guess you'd cry. I was pleased recently to see even Sun employee's gawk at some of the crazy policies ...
- Community hacker - You know, I don't dispute that specification can be important ... quite frankly, I don't see it worthwhile to sacrifice my time away from my family to try to write a spec, have to debug the basic macro in the template, try to get some help from Sun, only to find my plea for help silently ignored ...
- Sun hacker #1 - read your blog, pre-emptively setup a meeting, produced proposal for improving some aspect, mailed me about it for comment, CC'd relevant stakeholders [wow].
- Sun hacker #2 - This is absolutely inacceptable. ... there are a lot of complaints about work done by you and your colleagues that caused a lot of frustrating work in our QA and for our developers. ... I'm very disappointed. - from someone whose opinion I rather value.
In the light of #2, reading again, I note two things:
- QA/UE could get upset wrt. big-print / pictures. The point is simply - that there is no need to patronize these people, they are competant technical people. Simple, easily discoverable features (eg. the quick-starter) should not need a specification.
- Humor - it is clear that my daughter hasn't filed an issue yet, and this is rather an exaggeration, and of course some parts of Sun are extremely helpful and responsive, eg. the GSL team - and others, indeed wrt. timely patch review the situation is improving daily. My main complaint is the specification process - at best - it dramatically hinders lots of the diffuse 'polish' changes we need.
Up lateish, into town with babes to buy misc. bits. Finally repaired the toilet flush - the long-suffering wife pleased. Lunch, slept a little in the afternoon, poked mail.

As I was heading to bed last night, I discovered some wise advice from my (6 week) old daughter Myriam:

Lets hope she lives to see her patch get up-stream.
Up, poked mail. Set about disabling the quick-start feature in the up-stream build because it doesn't meet the demanding standards of beaurocratic process; at least if it's disabled, people can easily get / ship it, and we don't apparently have a fork, and Sun gets to write the spec. / suffer their own pain.
Relieved on returning to my desk to find the result of my query to the StarDivision library, a critical document (note the much-thumbed aspect). Since there is a bug in some OO.o printing interaction, it'll be good to read up on some of the background rational, work out if the behavior is already specified, where to adjust the specification, take some new screenshots and so on.

Some may call this process unnecessary - but without this specification (including pictures & big print) how could User Experience, or QA understand what the feature should do ? Or worse - imagine coming back to this area, and not knowing if a bug was even a bug, and whether to fix it. Remember: there is no process so effective that it can't be improved by adding steps and stakeholders, and there is nothing too trivial to be fully specified.
Got on with some real hacking for once, fixed a number of issues, cwsaddtask'd them, pasted the number & description into the commit message (dutifully). If we add enough process barriers, and consensus generation stages to the OO.o development process, perhaps we will (eventually) end up with something easy to use.
Fixed several other quick-starter issues, it starts to look rather polished, marked CWS 'Ready for QA'. Out to hire a dinner jacket for next week.
Got an encouraging mail, from a pro-active Sun employee that cares about improving this relationship - wonderful. Call with Intel.

Up early, let Julia lie-in. Dug at mail - poked at bugs, re-assigning those awaiting UI feedback to the UI team: what's a 6 month wait for comment between friends, side-tracked by a number of issues in this state.
Discovered -Bdirect support has been in Alpha5's OO.o for a while now, did a few warm start measurements on an ancient laptop:
```
before: 5611 5604 5606 5662   Avg: 5621
after:  4977 4883 4863 4881   Avg: 4901
```
It seems the redirect script is no longer necessary, since _ZThn symbols are not marked linkonce anymore: neat. On the other hand, the speedup is only 13% - but this is an old PIII, need to reproduce on newer hardware; installed Alpha5 on some newer H/W in the background, X broke during boot; bother, filed bugs.
Did an m187 build, started merging fixes to CWS gtkquickstart2 - got a number of things in.
I begin to wonder, given the Sun love for strangling the abnormally patient people who are trying to help them build an office suite, whether 'Open'Office should be renamed Office, or perhaps 'Sun'Office, or perhaps even 'Bow-before-the- specification-process' Office, perhaps there is some good acronym for that.
Dinner, hacked a little later - tried to get some dmesg output for Andi Kleen during my I/O related hangs:
```
echo 1 > /proc/sys/kernel/sysrq
echo "t" > /proc/sysrq-trigger
```

Up early, tried to poke German HR again, somewhat unresponsive. Finished the queue of individuals to thank / round up details with from my Nbg trip we employ a lot of good guys.
Got in touch with HR in the end, dug at a curious -Bdirect bug afflicting Dirk. Into Cambridge for a Doctor's app't that didn't materialize; extraordinary.
Dinner, worked late reviewing documentation, got into a nasty situation of needing to catch up lots of things, and still the up-stairs toilet flush is unrepaired since Germany; and no hard-core hacking for days now.

Poked XMS & NPP for good measure. Tested a new Win32 build - nice to see Tor's Cairo/Win32 port slide-showing nicely, and optional arg. support in calc.
Team meeting, getting rather long, lots of good stuff going on though, good to sync with Eric Ward at the end: our new QA person.
Discovered that Elluminate managed to very reliabley record 40 minutes of black/sience for my OO.o training presentation due to some mis-configuration. Clearly an amazingly useful feature to have, not only that, but the half-duplex sound support makes for really poor interactivity.
Worked late, E-mail and things.

Up early, to work; I/O discussion going interestingly, Olaf produced a beautiful test case / numbers to show madvise misbehaving vs. fadvise / manual mmap/poke: neat.
Long time on the phone with Florian. Lunch with Bruce & Anne - generated a nice sample of the I/O starvation problem hurting my desktop for the kernel guys. E-mail, left & right, performance bits, wrote LXF column.

Mini lie-in, off to NCC - spoke on the 9th commandment: giving false testimony, no incidences of spontaneous fossilisation in the congregation this week: survival of the fittest.
Lunch, slept a while, then bounced babies for some time so J. could sleep. DVD with babes in afternoon, fine dinner.
Gordon sermon on 1 Peter 1:3-9 - "Resurrection Joy" in the evening, bed.

Up early, J. into town to the market, worked on talk for tomorrow. Sue & Clive, Georgina & Stephanie (and multiple dogs) arrived.
Ate a large & lovely meal, wandered onto the race course - with H. on her bike, a racing day, the place packed with cars where we want to walk; bother. Back, played some guitar with Clive, and talked about this and that, very pleasant afternoon.
Worked more on talk in the evening, very tired.

Babes up at the crack of dawn, H.: But I think you were in Hamburger; N.: Ring Roses, Ring Roses; M.<smile> - good.
Removed battery from the new Toyota to put it on charge, since I also left the side-lights on by accident only last week it surely must be the car's fault. Supplying a hearing aid to each customer so they can detect the "you idiot you left your lights on" 'buzz' would be a useful innovation perhaps.
Followed Herr Hopf's advice and got a beautifully working, suspending, resuming, Xgl-ing laptop display driver: nice. Started to review & try to prioritize tasks, set off the Alpha 5 download.
Out for 3hr meeting with potential new church leaders.

Up early, breakfast with the team - to the office, poked at mail & bits - tried to call Florian. Met Lubos Lunak in person for the 1st time, a hero notable for his performance work across the board; lots to talk about. Interesting and positive meeting with Olaf Kirch.
Out for lunch with Voijtech & Lubos, more compiler/linker chew with Lubos, accidentally woke Mike Fabian - deadly ill with the lurgy; another call with Olaf. Chewed over X with Mathias Hopf, got my laptop doctored by Holger, demo'd Ricardo's yast2-gtk work to Stano. Bid 'bye to the lads, caught a taxi. Smooth trip home Stanstead traffic is light mid-week it seems.

Up, breakfast with Kelli, JP, Robert - taxi to the office, talked to Coolo, met all manner of interesting people. On to the yast2 team, met Stephan Hundhammer - had a great talk, chewed over lots of issues, out for lunch with the yast team.
Back, wandered the corridors, long and interesting talk with Michael Matz, good man; caught up with Kelli/Nat/JP/ Robert etc. out for a fine Italian meal in the evening, met a lot of Ste[ph|f]ans and got to know Vojtech a little. Back to the hotel, via the bar, to bed.

Up at 5am, sorry 'bye to all manner of beautiful sleeping creatures; drove to Stanstead. Hacked in the departure lounge - got a hacked vmplayer image to boot my hacked yast2 installer; nice. The German experience started with fizzy water by default on the flight.
Met up with Mike Fabian at the airport, headed out to the office, talked to Marco at length during the day, demo'd yast2-gtk a11y, etc. Managed to miss lots of missable meetings. Out in the evening for fine local cuisine. Got to know Kurt a little, great to see the team after a long gap, and see new faces.
Back to hotel, talked to team, stayed up very late talking to JP.

To work, fixed i18npool crasher in new system-icu patch. Dropped Thomas off at Cambridge station after some quick yast-gtk hackery. Submitted my yast2-control-center-gnome package to autobuild's tender ministrations.
Poked at ISO re-writing, cramfs, and all manner of similar goodness.

Up early, off to NCC, J. did creche, lunch, (over) slept, woke babes up, off to Chesterton Sport's Centre for baby's Korf Ball practice: a slightly nerve-wracking feeling being in a sport's centre again after so long; H. seemed to enjoy it.
Back, dinner, bed, Thomas arrived, talked a while, bed.

Up early, shower, kids to market, back, out to Andrew's birthday party (at NCC), lots of fine food, home, slept a little, out to Georg'es birthday party (at NCC) - lots of fun toys. Back, DVD watching with babes, bed early.

Poked at the win32 build - cygwin managed to hang forking it's own 'sh.exe' - embarassing, seems it's related to it's own mangled paths: /opt/ instead of /cygdrive/c/cygwin/opt or whatever.
Some code review, looks like we can save >1Mb from libsal by dlopening exotic text encoding translation tables on demand, at least on Unix where UTF-8 rules the roost.
Once again stymied by subversion's Wow man, anything should be possible attitude, that makes it really hard to do simple things easily; I want to recusively tag the current directory since I'm about to do something I might later regret; cvs tag Foo seems to equate to a multi-step process:
- Work out the branching strategy / location for this repository, branches can be 'anywhere' man.
- Check out the 'branches' directory (non-recusively -N otherwise presumably you bomb your disk) to somewhere.
- Now svn copy the source to the new location.
Oh - did I mention that you also can't simply list the tags that apply to a given sub-tree, you have to go groping around typing long URIs in repeatedly, and working out where people ended up copying them etc. SVN while extremely powerful seems highly unusable for the rather common branch/tag/list-tags operations. Hopefully manually writing the revision in the ChangeLog will suffice.
Reviewed Joerg's helpful comments on my configmgr refactor, giving some interesting pointers to futher savings. Hopefully if we can get the front & backend close enough to each other code-wise, we can understand the intermediate bit well enough to do something about it. Marveled at the charismatic pidgeon in the garden raising one wing in the rain (washing?).
Spent some time making a nice package for my slab based yast2 control center, poking pdb, autobuild etc.
Out to Newmarket Open Door to setup the Linux box, got the hub in, and (after blowing up 1 network card [ prolly one of my hand-made patch leads shorted & killed it, sigh ] ) managed to get all 4 machines talking & connection sharing: yast is rather good for that.
Back for a team jolly that didn't arrive. Dinner, Coleen & kids playing around the table. Back to work to catch up lost time.
Dug at the textenc code - it seems there is an easy 1.7Mb to save from 'sal' by ripping the (almost never used) textenc (encoding translation) code out into a dynamically loadable library; for those who still have IBM_850 encoded data. While doing this produced a nice list of used textencodings on startup/edit, very surprised to discover several (low number) encodings still in very active use: turns out to be a vicious polymorphism related problem. Thank goodness most low enum value encodings look sane in the ASCII range.

To work, poked at mail, updated some bugs - with screenshots to beautify. Started doing a number of timing tests, it seems Linux is doing better than I expected, at least on my Desktop machine. Took a lot of measurements, drew some graphs etc.
It's pretty astounding that OO.o on Win32 takes 13 seconds to cold start, where MSWord takes ~3 seconds; in fact Word can cold-start quicker than we can warm-start, a little depressing. I wonder if any good (free) profiling tools exist on Win32, to see if there is something particularly dim we're doing, that we could avoid.
Knocked off for dinner, putting the babes to bed, good to see Janine & kids, dinner with Mum. An hour of training NTS on OO.o 8:30 onwards, J. out at a Bible study.

Up early, dealt with babes while J. slept; to work, poked mail, posted action items. Back to configmgr - managed to axe memory::Allocator completely, then after another chunk of more careful, intricate tedium - removed the 'Pointer' class.
Posted my configmgr re-factor patch to Stefan B. hopefully he likes it, mail with various likely suspects, some progress wrt. up-streaming this & that.

Up early, read mail - priceless comment from a (much loved) colleague about the OO.o QA test-tool's promiscuity: ... the testsuite is very slow (many sleeps around)... (3 days to get a result).
Filed expenses at some considerable length, eventually got them done, missed the 60day deadline for 1/2 of the last lot though (hmm). Back to configmgr, now crashing - apparently removed too much too fast this time.
Team meeting, good work being done left & right. Poked people about switching to Google SVN & getting gmail accounts setup. Call with Tor. Petr did some great QA on the quickstarter CWS, found several problems and nominated it.
Ripped more out of configmgr: simpleheap, HeapManager, Heap gone, loads of confusing typedef / renamed types gone, started the attach on Allocator.

Up early, fed babes, to work. Mail, poked Stefan, wrote up some thoughts, bits of admin. Lunch, more mail, urgh. Grandma arrived at 3pm or so. Hacked at configmgr - bits flying removed the custom allocator thing, no problems so far (beyond valgrinding OO.o causing my H/W to lock up (hard to file a bug report)). Ripped more guts out of configmgr with no ill effects, neat.

Up early, breakfast, packed things, drove home.
Ben Elvidge over for lunch, lovely to see him again and catch up with goings on Nr. Oxford, lugged some shelving being slung out from Nigel & Caroline's house together. Watched some DVDs while J. tried (abhortively) to sleep, Tea early babes to sleep.
Tried to get to bed early, thwarted by Myriam who seems intent on complaining about wind/hunger she doesn't have, poor wife over-wrought.

Up early, left the H. & N. with the Grandparents, set off for Coventry - drove, drove some more, more driving - arrived in time for the service. Lovely wedding, beautiful bride & lots of old common friends.
On to the reception at the Coventry Guild-Hall. An incredible building, steeped in history, fine food, good company, very enjoyable. Myriam very good, lots of sleeping and little complaint.
Talked much of the evening, then set off back home, drove & drove late at night, bed rather late.

Up early, suprised to discover the wife had been reading 'Peopleware' (even I could understand it), though I hadn't got to it yet, no Economist - so started; interesting so far.
Poked at mail, amused on analysing an OO.o corrupted my spreadsheet bug to discover a worrying preponderance of '0x0d 0x0a' bytes, including the 2 terminating ones, a total lack of 0x00's and a pre-ended " character; looks like some ninny loaded it in some DOS text editor, then saved it again. Sadly although some files in OO.o cough types.rdb could use a similar treatment, it doesn't look possible to automate a 0x00 re-insertion to restore the content.
Long phone call with Kai. Squeezed more memory, code and performance out of the configmgr, analysed the duplicated strings again; framework burning > 1/4 of the reduncancy now (tons of ".uno:SaveAs" type entries). The reason we have 145 instances of 'ALIEN' unfortunately turns out to be poor representation of the binary config cache, shame.

Up early, dealt with babes, work lateish. Poked at OpenSUSE10.2A4 - filed 10 bugs around the desktop, most of them sillies, noticed a few more while responding to queries.
Drove to Bruce & Anne's, lunch. More mail, poked at go-oo.org, running out of space. BU employee call. Committed / enabled my sc-optional work, and poked Eike/Jody for more review.
Committed the libegg pieces for the systray quickstarter, updated the CWS, though didn't re-sync it (too painful ?)

To work, phone interview on time based releases. Interesting to see a new Sun GullFOSS aggregator thing spring up.
Hacked at calc all day - talked to Eike who seems to accept the approach, which is great. More digging at that, poked at some StarBasic profiles with Noel - we need a ~2 order of magnitude speed win in there at least, 3 in some areas - should be fun.
Coleen, Anna & Ethan around for dinner, good to see them. Cell group in the evening, last evening with Thomas around - sad really. Bed late.

Up early, packed the girls off to 'pre-school' (ne play-school), and got to work. Vexed by my ethernet connection, clearly buying the cheapest cable crimper on offer wasn't the best plan: badly formed cable bent one of the 8 springy pins inside the expensive laptop: check your cable before use.
Interesting phone call with KaiB. Talked to travel agent, booked Nuremburg trip, looking forward to meeting some of the SuSE hackers I've worked with in person.
Super-long team meeting, lots of good work going on. Built DVD ISO from OpenSUSE10.2Alpha4, set off install.

Up early, poked E-mail, nothing much - hacked at optional argument patch, beggining to look quite sweet - created a fairly elaborate regression test sheet, the back-compat piece is looking good; saving & re-loading ~everything works; excellent, now for the core hack.
Out for a run with J. - lovely to talk to have her to myself for a while, T. left in charge of the babes, with DVD to assist. Back somewhat later, babes to bed (bible story, prayers, children's story, sleep), dinner. Watched a Vet movie, sleep.

Up early, fed babes, off to NCC - Daniel speaking briefly; Ron encouraged us at the end; had DT, Zoe & Jackie back for lunch, chatted for much of the afternoon.
Slugged, Tea, bed, Thomas back, burped babies, bed again.

Up early, off to the market, back - potatoe peeling at length. People started to arrive: James, Kate, Adam & Laura in the 1st batch, Christian, then Thomas. Badmington, Ham, Egg & Chips, fine meal, good conversation.
Lost everyone after a while, rather early, slugged, read to babes, washed up, bath, a little hacking afterwards.
A while ago (ok so several years) I read Eric Raymond's interesting testimony (an excerpt): That day I was reborn; from a skinny lame kid with a flute into a shaman and a vessel of the Goat-Foot God, the Piper at the Gates of Dawn, the Horned Lord.. At the time I thought how tendentious some of it was, but clearly 2 can play at that game; so eventually I got around to creating a less elegant reply Why I am a Christian, with some of the answers to questions I wanted people to ask, but they didn't. Still need to finish up my random selection of anecdotes wrt. other worldviews at the end; it took a long time, but this hacking stuff has been preoccupying me.
Hacked a little at the detection of missing opcodes in ScTokenArrays (seems to be a flattened / parsed form of a formula). Got somewhere - missing args detected, re-written & substituted with a string.

Up early, listened to Thorsten's OO.o to XCanvas talk while reading mail. Poked at a nice interop problem: optional (mid-list) arguments in calc, the 'PMT' function isn't happy; quick chat with Jody.
Tried to come up with a nice generic way to share function / argument data between ODF export and the calc core wrt. missing arguments: sadly rats lurk here. Switched to a more pedestrian approach.
Late night VBA call, went ok. Dinner, back to creating nice missing optional argument test case, reported silly bug in the ECMA formula spec.

Mail thrash, it seems IBM are interested in helping get LWP files importing into OO.o, our efforts to reverse engineer the format are somewhat stymied.
Was recently provoked by a criticism of the LGPL into noticing -just- how successful it is, particularly among large companies. ~Every application that runs on Linux links to the LGPL glibc (it's hard to do anything without that) so IBM's DB2, or Oracle's D/B would be notable here. Also, the number of interesting companies shipping code licensed under the LGPL is large (some examples): Microsoft [Interix SDK], Sun [Gnome], Sony [NSP1], Sharp [Zaurus], Nokia [770], Novell [OpenSuSE], Netgear [my router] & Tivo (of course anyone shipping a Linux based appliance / device does), gave up extending the list, doubtless there are a lots more.
Poked at the Novell Developer Demotivation Tool (the innerweb provides a number of such things), unsuccessfully. Then Kelli's staff, then onto FATE - finally used the thing in anger: hmm, mailed a set of queries off to Klaus & nursed my wrist.
Hannah now scared of 'tomatoes (with faces)' under the bed, a progression from the dragon in the shower-room, to 'Skull and Cross-bones' images (amazingly prevalent), to Scooby Doo [ having never seen it ].
Dinner, hair-cut from the wife: "I'm sorry darling I just assumed you had put the no. 3 on the clippers, you normally do" - an amusingly severe/cold bit at the back of the head. Bed early.

Up serially in the night, N. found twice asleep on the landing (in crawling position): strange, the dire warnings about not waking us clearly had some effect, though not of keeping her in bed, poor dear.
Call with Matt McKenzie 1st thing, poked mail. Poked NTS wrt. training again. Listened to Jody's talk - great stuff, also Thorsten Bosbach's testtool talk. Concurrently poked at a curious pivot issue. Matthias Bauer's talk also great.
Poked at some Visio OLE2 embedded in .odt on Win32 issue that Fridrich is in hot persuit of. Suffered Win32 unwillingly, I thought I had Visio installed, but couldn't find it. Dug out the DVD, ran setup "upgrade/remove" - apparently it's installed. Eventually found it languishing on the desktop (one of 45 icons), apparently putting it there removed it from the 'All Programs' menu (50 items at the root of the tree). Yearned for the searchable, Novell app-browser on top of Win32.
Customer conf-call, chat to Kelli & Rob.

Up every hour in the night with N. seems to be finally objecting to being usurped as the smallest & most needy; tiring.
More mail, action items, write-ups, blah. Upgraded cygwin (a leap into the dark at the best of times), perhaps that will cure the plague of crashed shells. Conference calls.

Up early, looked after babes while J. slept. You know something is being learned when N. wakes up her mother with Hello Dorgeous. To the mail mountain, hundreds of unread personal messages, urk; hard to resign oneself to a ~week of non-hacking.
Nice to see the new OO.o build-bot having an IRC presence. By googling [ couldn't see a link on the OOoCon official page (go Marketing !) ] managed to find the conference videos.

My hot picks (snarfed from the nice Kiberpipa site).

Spreadsheet Compatibility. Forwards, Backwards, and Sideways

Jody Goldberg (Novell)

12.09.2006

Play video in browser

Play: Video Audio

Download: Video Audio

Lightning Fast Startup

Michael Meeks & Jan Holesovsky (Novell)

12.9.2006

Play video in browser

Play: Video Audio

Download: Video Audio

GStreamer in OpenOffice.org?

Cedric Bosdonnat & Radek Doulik (Indep. & Novell)

13.9.2006

Play video in browser

Play: Video Audio

Download: Video Audio

Building OO.o (faster?) with Jam

Kai Backman (Google)

13.9.2006

Play video in browser

Play: Video Audio

Download: Video Audio

Modular Building, splitting the build

Caolan McNamara (Red Hat)

12.09.2006

Play video in browser

Play: Video Audio

Download: Video Audio

OpenOffice.org 2.x and beyond

Michael Bemmer (Sun Microsystems)

12.9.2006

Play video in browser

Play: Video Audio

Download: Video Audio

The Novell migration to OpenOffice.org

Debra Anderson (Novell) [ starts +9 minutes ]

12.9.2006

Play video in browser

Play: Video Audio

Download: Video Audio

Watched Michael Bemmer's talk: very interesting, it's great to hear StarDivision emphasising reducing Barriers to entry the priority of building community etc. Pleased with Sun's current direction. Couldn't find Noel's VBA talk, or Fridrich's - both rather good. For those with hyper-low bandwidth, you can get slides.
Tried to fix the HTML to not make the planet barf. Got mired in some mess as to where pubDates should go & so on; utterly evil. Desparately quested for a decent DTD link for RSS and got no-where, my (apparently) valid RSS is unworkable, found this urk; and (apparently) no DTD too; so xmllint --valid can no longer save me; grief. Talked to jdub: Atom is (apparently) the sane format here, and I need to use pyblosxom's static publishing.
Cached thumbnails on gnome.org & switched 1/2 the video links to another server; thanks Laurent. Knocked up column for LXF on the cusp of missing the deadline, generated more copy for Mehran, urk.
Bought some books on S/W management, lots to learn no doubt.

Up early, read to children; out to NCC: speaker from The Leprosy Mission, on their work & why they do it.
Back for lunch, slept much of the afternoon, dinner, children playing with present of dressing up clothes, including the dangerous favorite: high heeled shoes. Bed early.

Read to children, played with them, nondescript day(?)

Up late, breakfast, some mail poking, very tired, a little work while babes watched some DVDs; FTO. Interesting phone call, started to write up some notes.

Breakfast with the lads, off to the university - Cedric very kindly provided us with a room; team meeting much of the morning.
Into town, lunch, bid 'bye, to the flight; Zaheda & John on the same flight again, chewed over some of the conference developments. A little more hacking, saved more configmgr strings & memory with a more intelligent cache.
Swift passage through Stanstead - although difficult to get to the pick-up point; J. & crew picked me up - wonderful to see her again.

Up rather too early, off the the ODF keynote / panel thing. Fairly interesting, talked to Rob afterwards, nice to get a commitment to help OO.o with unwinding the lwp file format.
Caught Michael Brauer's talk on ODF, interesting. Missed lunch while chewing over the state of the world from IBM's perspective, with Bob and [fixme]. Got a nice demo of the Lotus Workplace client, which looks interesting.
Customer meeting, saw Fridrich's nice talk on building community (from a developer).

To the conference, Debra's keynote went well. Lots of ninja lizard video action. Panic setting up laptop for Noel's presentation: burned 30mins. Back to back to back interviews, 1/2 a lunch with journalists, back to my startup talk (5mins late). More interviews, finally finished & bid goodbye to the nice PR folk.
Caught some of the chart2 and Caolan's talk, interesting, modularising OO.o at the build level may solve many of our buildability problems.
On by coach to the city of Lyon hall, beautiful art, architecture, fine canope's, good company - lots achieved. Kebab afterwards with the lads, tram ride to the hotel, and a pleasant drink with doko / the 3bclean lass.

Breakfast later than expected, laptop-time issue. Walked to the conference, sculled arou