Stuff Michael Meeks is doing
|
|
This is my (in)activity log. You might like to visit my employer
Novell which is an amazing company, and also
Dell who in days of yore provided me with a
free laptop for Gnome development / conferences.
Also if you have the time to read this sort of stuff you could enlighten
yourself by going to Unraveling Wittgenstein's net or if
you are feeling objectionable perhaps here.
Older items:
2008: (
J
F
M
A
M
J
J
A
),
2007: (
J
F
M
A
M
J
J
A
S
O
N
D
),
2006,
2005,
2004,
2003,
2002,
2001,
2000,
1999,
legacy html
- Normal LPO day, swam etc. Out in the evening with some
lads for a beer, good to get to know Ben & Brian & Scott -
fun; bed rather late.
- Up lateish - lie in. N. insisting on calling Croissants
'Crussles' in a very sweet way. Creche, continued study on Jude:
snatching things from the flames.
- H. dragged unwillingly from the FFT for bed, having
apparently spent most of the time in a castle 'tent'.
- Out to Noirmoutier in the afternoon, a ride on a two-story
horse-roundaboutey thing, crepes in a fine cafe. On to see a real castle
- answered lots of questions relating to princesses (why are there none
in France ? etc.) - an entirely new way to see castles from a young-female
perspective.
- Call with Florian in the evening.
- Up, creche while J. swam. J. & H. shopping. Swam in the
afternoon at some considerable length. JP called wrt. planet.go-oo.org
migration, great work, though woke the wife.
- Up, kids to play, study, kids to bed. Met Allison &
Stephen, off to a local lake for sand-castle building; met Graham &
Sarah from Sheffield.
- Back, cleaned (already dirty) house, chips, bedtime story,
relaxed & read the Economist, while trying to get flights re-arranged,
etc. for business on Monday. Bed.
- Up early, fed babe. Croissant for breakfast, this is the life.
Took the babes to play at 'Sticky Fingers' (swiftly renamed the 'Funky
Fun Tent'(FFT) for extra oomph). Bible Study - turns out it is 'food week'
too. The littlest back to bed, went shopping.
- Lunch, lovely French food etc. Swam in the pool - somewhat
cold, but wonderful flume, nearly lost both babes on their solo journey
down it; N. very proud of herself: "I sunk like a stone !".
- Then off to the playground with the babes. Call with Carlos
turns out it was a mistake bringing the phone on holiday. Back, chicken
curry dinner. Worship in the evening, bedtime story-time for kids at 7pm,
bed.
- More packing, exciting Airport experience for the
children (and at a human hour): Stanstead to Nantes - met an interesting
linguist chap 'Pilippe Patr' on the plane, had taught much of the French
football team French apparently.
- Queued for car, met another family also on the way to La
Pas Opton, got there eventually. Neat, but complete mobile home to stay
in. Terrible night - Myriam screaming much of it. Text message from
Carlos - need to fly to San Francisco next Monday; hmm.
- Up early, gave talk at NCC on historicity of the NT,
back home - started packing.
- Kept waking J. in the night, trying to re-factor her
into an UNO-ized version: clearly time for a vacation. Up late,
Off to NCC - Simon Matthews holiday slide-show, then Bob - a
partner at Morgan Lewis:
a Christian and expert in International Arbitration spoke on resolving
disputes. Extremely interesting & thought provoking, illuminated
with anecdotes from past cases; and with a sample role-play.
He recommended the Centre for Effective Dispute Resolution CEDR.
- Back for lunch; played with babes; set about more
research / preparation for talk tomorrow. Worked until rather
late on this. Left fixing up my evil proprietary driver ati
problems until midnight: not good.
- Up, drove home, chat with Michael; tried to tidy loose
ends up for my vacation next week. Call with Mary & Kelley - great
work. Wrote my LXF column. Chased & found ORBit2 issue for Jules.
- Finally back to layout again: nice, hacked away cleaning
up this & that. A little more work and the infrastructure to port
the zoom dialog with ~5 lines of code-change (and a new xml UI
description of course) is there. Of course, the wrapper API will need
expanding: incremental work. One of the nicest things about the design
(beyond making it possible at all), is (perhaps) the ability to
prototype this early in ooo-build, and maintain a huge large set of
scattered (but small) changes across the code-base (at least that is
one idea).
- Wrote up notes for Ricardo & Kohei, looking forward to
getting back and seeing what they've got up to. Worked rather late.
- Poked mail, Quick call with mba, pushed Ricardo's latest
yast2-gtk work through autobuild. Dug at another interesting ORBit2
bug: turned out a pointer had "Do you w" in it instead of something
sane: the joys of 64bit pointers; clearly some valgrinding of the
gnome-panel is called for.
- Call with Noel. Generated a list of cool hacks. Quick
call with Carsten, filed my awt button issue.
- Drove south to the parent's. J. has to attend Louis'
book-launch tonight. Poked JP, mail, back to layout. Worked on
sermon in the evening.
- Slept badly - too hot. Dug at mail. Did some polish
& testing on yast2-gtk, and released it. Lunch.
- Much more layout hacking, going well until the thought of
creating a UNO service from the thing came to mind: plunged down a
multi-hour rat-hole of obscure, tangled interfaces. Sure, you can
bootstrap from an installed OO.o - but, not if you want to have a
new custom component. Of course, you can try to register it, but not
if the service factory is read-only; cue useless LXR grokking for
generic interfaces, fruitless reading of 'stoc' code etc. ended
up wrapping XMultiServiceFactory with some hack that does what
is necessary. Why there are ~no standalone unit tests for any
high-level bits of OO.o - now you know.
- Then of course, the attempt to use the 'Universal
Content Broker' also completely wedges up, can't grok a "file:///"
URL, yet more time burned reading through ucbhelper, ucb etc.
futile: best to cut/paste something simple & usable like
xmlscript's xml_byteseq that 'just works'.
- Finally got the wrapper API connected up via widget
names to the awt peers constructed from the XML - good. Started
to uncover lots of evil in the toolkit / vcl interaction:
Buttons fire clicked events, unless they are 'special' buttons:
OKButton, or 'CancelButton' when they don't, but 'HelpButton'
does - curious.
- Dinner, tiny cell-group, call with Zaheda.
- Back to work, plenty of mail to sift through.
Chat with Caolan, more mail. Reviewed Jules' nice ORBit2 timeout
re-work. Watched Linus' git talk. Dug at a layout VCL/AWT API
prototype. Call with Stefan Taxhet.
- Lunch, team meeting. Hacked away at the now infamous
'zoom' dialog, using it to initially populate my VCL-like wrapper
API. Mostly hiding the UNO invocations to avoid exposing exception
bloat to the callers; should profile that though.
- Tony, Janice & Lydia around for dinner, up late.
- Up disgracefully late, 30 today; breakfast, present
opening: lots of wonderful things. Lunch, sound-system arrived
from Robert; fun. Off to the Funky Fun House for crazy
rushing around child-mayhem.
- Back, fed kids & put them to bed. Out with the
wonderful wife for a drink & Indian meal, lovely.
- Up earlyish, cleaned up the grisly aftermath. Off to
All Saints - creche etc. On to Emily's baptism at the school
outdoor swimming pool. Rather a brief ceremony, lunch afterwards.
- Home, tired children (and adults) demanding their movie
"on the wall", obliged. Bed early.
- Up early, off to the market with Myriam, got stocks in,
observed horrifying relational nightmare in Argos; complete with "I
hate you" chanting female & unpleasant baby-carrying male;
prayed hard.
- Back, frantic house cleaning, food preparation: 30'th
birthday party today. Family & friends started arriving - pleased
to see so many people: the Atkinsons (pregnant), Hendrys (likewise),
Brighteys, even the Ryes from Newbury, lots of Meeks, Heathcote,
Williams, Morgans, Webbs, Braven-Giles, etc. Fine to see the children
gambling in the garden, talk to so many old friends. Badmington in
the garden until a deluge & ad-hoc water-fight; hmm.
- DIY Pizza party in the evening - went rather well,
watched The Princess Bride again, talked until late, bed.
- Up early, used prime hash table sizes instead of powers
of 2, interestingly no siginificant speed win - perhaps our hash is
too good.
- Continued profiling - the store code looks increasingly
suspicious as a source of inefficiency: 40% of OO.o's locking on
startup is done there. Knocked up a simple patch to cut that to 25%
hopefully Matthias can improve it yet further in his re-work.
- Nailed the irritating gnome-screensaver crash-on-start
problem; an underlying bug in XGetClassHint's error handling - worked
around it.
- Spent some time writing a report, status report, conf-call.
Dinner watched movie with babies, finished & sent my report.
- Poked Josef & co. wrt. my cachegrind bug, too useful
a tool to be without. Dug at startup profiling some more.
- Stared at kcachegrind output for some hours; discovered
a few rather silly things: 3% of startup time calculating CRCs on
data that will always be correct (or your hard-disk failed). Also,
improved the string intern hash (binning the stl hash_set) - saved
170k and 30% faster hash, and 1% faster startup.
- Up, chatted to Hubert - the insomniac: the excitement of
hacking on OO.o is clearly getting to him. Back to binutils. Spent
some time chasing an ImageList related crasher, nailed the only issue
I could reproduce.
- Lunch with Lydia, looked at the pull-up to m215, getting
harder now more things are going up-stream, too difficult for now;
urgh; back to binutils. More profiling, showing no improvement - fun.
- Caught gnome-screensaver misbehaving, and updated bug.
- Call with Kelley, Nat etc. out to Alba's for dinner with
the children, met Jackie. Back for cell-group later; couldn't sleep.
- Pleased to measure OO.o startup on SL10.3 and see a warm
start time of 850ms or so; nice. Apparently (from 2 runs), -Bdirect
is now (in pseudo-cycles) only a 2% additional win, or 6% on a Geode.
- Boggled at the collab.net animated
advert on slashdot.org: "Want to jump of your version
control tool? Get Subversion supported by collab.net". Where
should we get off ?
- Did some ORBit2 review, to re-factor the new connection
timeout support that seemed to get into 2.14.4 (should pay more
attention there clearly).
- Banged on binutils all afternoon, with surprisingly little
joy, bother.
- Not much mail: nice, looked at some bugs. Started to poke
at cleaning up some old platform issues: got an OO.o build running on
SL10.3, the best linker test case. Played with Ricardo's latest
yast2-gtk installer: looking a lot better, fixed a few issues &
improved it a little.
- Burned the CPU building OO.o, while running callgrind on
another OO.o, and working on glibc. Positive ESC meeting.
- Up in the night, N. having nightmares, up early, dealt
with babes, NCC, creche, back for lunch with Katey, DT & Zoe.
Discovered the automagic oven hadn't turned on: lots of un-cooked
lamb.
- J. whipped up a lovely bacon carbonara instead, nice.
Lazed & talked all afternoon most pleasantly, DVD 'Sliding Doors'
in the evening: curious if risque.
- Up disgracefully late; poor wife bore the brunt of the
labour. Lunch, off to NCC to clean with little babes, while J. took
H. to a party. Hoovering technique improving.
- Long walk back, in very hot sun - saw lots of horses;
ice-lollys and paddling pool in the garden, bed early.
- Extracted the patches necessary for the layout prototype
into ooo-build HEAD to make it easier for people to play with.
- Up late, read mail. Lunch, worked on documents. Sorted out
the yast2-gtk dump files to try to get them into .cz past whatever mail
filter is there.
- 2 hour PRD review, some fun moments but fighting for breath
at end: layout hacking to revive - dunged out some cruft,
- Up early, a day of phone calls, got some layout stuff done
and synched with Ricardo: we now have both v & h boxes, and the
canonical v/hbox test works nicely: except for the unfortunate toolkit
XLayoutConstrains impl.s, that appears to revel in picking sizes that
are too small.
- Submitted (later) OOoCon paper on layout (with Ricardo).
Call(s) with Kelli - good to have her back. Another interview. Fixed
the vcl button CalcMinimumSize method, and got a nice screenshot
finally (of course this is Ricardo's work too) the archetypal vbox/hbox
with permutations of expand & fill settings:
- Of course, the really fun problem is not the easy bit of
creating the layout, or the XML parser, but making it possible to
quickly & non-invasively retro-fit existing code.
- Very late night call; timezones: they suck.
- Dealt with mail, call , back to the layout hackery.
Lunch with Lydia, more layout hackery, poked Nat, call with Patrick. More
layout work - qiute exciting, finally got something working; very few lines
of code still.
- Poked mail, wrote hard-to-write document, etc. Hacked away at
layout - binned most of the old/ugly toolkit re-work, nice. Split the code
out into a new top-level module, inspired by Christian's work - with it's
own IDL build to save offapi / offuh faff.
- Lunch, team meeting, more layout hacking, meeting setup -
starting to be quite pleased wrt. prototype; still no layout of course -
but fairly hopeful code so far.
- Tony & Simon over in the evening, talked Cell planning
until late.
- Up early, poked Nat. Call with Jeff, lunch, PRD review with
JP, Guy etc.
- Had an evil problem with a git repository: if the version you
push to has another git repo inside it (ie. a .git directory) very odd
things happen: updates.
- Helped Ricardo with various bits - a little more layout
hackerage, getting somewhere at last.
- Up early, N. still sick (can't hold down water), J. not
feeling good either. Off to Church with Hannah. Tony speaking on
Jesus in the bible.
- Back, lunch, sunned ourselves in the garden, played with
babes as N. / J. slept. Watched some of The Slipper and the Rose
with babes in evening.
- Great talk by Gordon
What the World forgets on the Genesis flood:
- wrt. dating of the flood: only 1 piece of biblical
evidence: genealogy - Gen 11
Noah to Abraham: 10 generations. Abraham generally agreed ~2000BC. Add
Gen 11: ~2400BC - a possibility: assuming the genealogy is exhaustive
& complete.
- Digs at Ur in S. Mesopotamia, initially excited Christian
world: found siltation layer: right at spot - 2400BC: little/thin layer.
Doesn't correspond with adjacent cities. 1000 years before - big siltation
layer: 20feet thick.
- Leads one to suspect - genealogies highly compressed
Matthew 1:1
Jesus, son of David, son of Abraham - covers 2000 years.
Can't impose our cultural preference for exhaustiveness on the text.
Several other examples Luke 3
has Cainan before Shela cf.
Gen 11:12
'became the father of' etc.
- In conclusion, cannot safely deduce much except, can't possibly
be later than 2400BC. Perhaps much earlier, vast amounts of time ? time of
last ice-age 50kBC perhaps ?
- Sir
Leonard Woolley - dug at other cities - no corresponded floods at other
cities at same dates; even for the 20foot strata: big flood-plain of
Tigris / Euphrates.
- Other issues:
Jericho - archeologists: uninterrupted occupation there, back to 7000BC,
no evidence of a flood at any point. Bible-believeing, sincere, careful
archaeologists - pay attention, say flood at least 7000BC. Gordon's guess:
50k BC, it's a free country, anyone can guess. Time before anyone migrated
to Australia, or to the Americas, Brazil ~30k BC.
- How extensive - less dogmatic: more honest, what bible
actually says. One thing non-negotiable: everyone killed, all mankind
wiped out - for sure. Does that mean - the flood-waters innundated
Antartica ? necessary to conclude, entire sphere - if people not living
there yet.
- Text calls - covering the 'earth' - hebrew word: 'land' /
'region'. Similar passage, later in Gen 41
And all the countries came to Egypt to buy grain from Joseph,
because the famine was severe in all the world. No bible-believers
postulate the Eskimos or Australians showing up: context matters.
"World War II" eg. : tons of nations not involved; can't be sure from
the text.
- Also, practical problems: worth acknowledging, becuase
Christians know an enemy that is a deciever so cultivate a critical mind:
examine the facts. Difficulties.
- Could every animal fit in the ark plus food for a year ?
How did they get there ? did the Marsupials swim the Indian ocean (if
so, why bother with an ark ?), and did they all swim back only to
Australia afterwards ? What about the fresh-water fish ? was there an
acquarium on-board ?
- Evidence of the whole Bible: another biblial text -
a parallel description of creation talking of the
primeval waters says of creation You set a boundary they cannot
cross; never again will they cover the earth.
- God was angry at people, not Pandas. Further evidence: it
seems to be an eye-witness account: Noah sent out dove, it could not
find a place to set it's feet. Did Noah's Dove circumnavigate the
globe ? checked Antartica / Everest ? seems unlikely.
- On the other hand - clearly, a vast, unprecedented,
universal flood. Geological evidence - inconclusive, nothing more
recent than 7000BC, possibly earlier. Some interest: a Columbia University
study on the
Black Sea sees a 60k square mile area innundated in a matter of months
for example.
- Was there a flood ? you bet there was, let God be true &
every man a liar. Not only the testimony of scripture but external corroboration
too. No legend more universally attested than the legend of a flood
that wiped out all of mankind before the birth of history: 300 narritives,
Persians, Indians, Chinese, Japanese, American Indians, Australians, Greeks,
Egyptians, Babylonians, Sumerians. etc. vast literature. Some of them -
hallmarks of 2ndary invention: missionaries got there before the anthropologists
Earliest accounts: written in ancient times: remarkably parallel to biblical
account.
- Commonality - all other accounts, blame God: it's his fault:
capricious, or annoyed - the human vermin make too much noise etc. Not so
the Bible, Gen 6:2 - biggest foretaste of final judgement yet to come.
- Why ? man was destroying the value of life, and trashing the
value of marriage: Lamech does both. Luke 17: "Just as it was in the days
of Noah, so also will it be in the days of the Son of Man. People were eating,
drinking, marrying and being given in marriage up to the day Noah entered the
ark. Then the flood came and destroyed them all."
- Gensis 6 - people corrupt, violence: people imposing their will
on one another, explotation - it's divorcing your wife, because she is
inconveient: provokes God. Rampant sexual immorality - offends God.
- There were many in Noah's day this is it, grab all we
can get or What I sow, I don't necessarily have to reap. Did I
live my life 'without regrets ?' The only reason so many have a clear concience
is that they have a lousy memory; forgotten grevious harm done to fellow man.
A personal God, takes personally our offences.
- Ironic twist - men become their own 'sons of God', and God does
the repenting: repented the Lord that he had made man on the earth
His heart was filled with pain Gen 6:6. It grieves God, he repents -
and takes back the gift of life.
- God's patience is no evidence of indifference: waters of baptism
an echo of the flood:
1 Peter. Jesus' baptism similarities: God echoed what he said of Naoh
with whom I'm well pleased, the Holy Spirit as a dove - returning to
the ark: place of safety in the flood - hide in Christ.
- The Ark - not a boat: no bow, stern, rudder, no oars, no sail.
A box, a temple - tripartite structure: same shape, virtually same dimensions:
a temple - clean & unclean separated. Jesus' body also a temple of the
Holy Spirit etc.
- Building an ark 100 years before the flood looks like a waste
of time
2 Peter 3 they deliberately forget ... the world of that time was
deluged & destroyed do not forget the Lord is patient not wanting
anyone to perish, but all to come to repentance.
- N. sick repeatedly in the night - nasty; poor dear. Off to
Whipsnade Zoo - a
lovely hot day. Saw lots of fine animals with the babes, talked to Sue
& Clive.
- Drove home late afternoon, dinner, bed - J. not so well too.
Hacked control construction up in bed: got a nice, simple,
& clean XML representation going for awt widgets: good.
- Chewed mail, documental digging. Tried to work out why
the tiny xmlscript/test test takes such an infernally long time (a
handful of seconds) to start on a really fast machine: just to render
a hand-full of simple widgets. It turned out to be doing some impossibly
expensive operation on all fonts to satisfy ImplCalcMapResolution,
clearly a missing cache somewhere.
- Call with Christian Lipka - nice man; chewed over the problem
space; set to hacking something pleasant wrt. layout, burned away. Set
off for Sue & Clive's over lunch, got there.
- Poked again at the code, grief the awt appears to export only
a crippled get/set property interface instead of an introspectable
XPropertySet; how odd.
- Lovely dinner in the evening with Sue & Clive, bed late.
In case it's not painfully obvious: the reflections reflected here are my
own; mine, all mine ! and don't reflect the views of Novell, The
Lithuanian Gov't or Arnold Schwarzenegger. It's also important to
realise that I'm not in on the Swedish Conspiracy.
Occasionally people ask for formal photos for conferences,
bio.
or fun.
Michael Meeks (michael.meeks@novell.com)
