Older blog entries for lkcl (starting at number 179)

21 May 2003 (updated 21 May 2003 at 14:25 UTC) »

i very much appreciate it when people such as yourself make comments that bring me back down to earth.

i should point out that my anger - even after three years have passed - is directed quite specifically, and not at "open source" in general. it's very interesting to note someone else's recent diary entry on here about commercial exploitation of open source leads them to encourage others to rethink their relationship and attitude to open source.

perhaps my experiences - which are at the extreme end of "bad" - will encourage people to, i dunno, set up a guild of open source programmers; a trade union; a charter which outlines the expected relationship between programmers and potential employees; even a freemason's lodge - heck, i don't give a monkey's as long as it makes sure people don't end up going through the same shit that i did.

so, without going into too much detail - and answering your question directly at the risk of inciting wrath:

i put three, maybe four intense years of my life into samba, a major project that has won awards and attracted commercial sponsorship from several companies such as SGI, HP, IBM, VALinux, TurboLinux, Linuxcare;

during the dotcom boom, NOT ONE of the linux companies invited me to take part in their IPOs, whilst at the same time, one of those linux companies invited the brother of one of the samba developers to take part in its IPO;

the so-called samba team leaders were ALWAYs jealous and incapable of understanding how much work i did and how far i was pushing the boundaries of interoperability with NT, and they couldn't handle it; now they are running into design difficulties because they were incapable of listening to my advice;

psychological and financial pressures made it very _very_ difficult for me to communicate effectively and also to be able to negotiate properly, to the point where one of the samba team developers not only wouldn't look at any code i wrote but also would rubbish literally _any_ ideas that i had, even if they had successful usage and grounding in other well-known and well-established projects.

basically what happened was a massive failure in communications and relations that would, and i am not joking, have led many people in the same situation to a nervous breakdown.

to cap it all, my employment contract with linuxcare required that they own any intellectual property rights in code that i wrote and ideas that i had (so i made sure i didn't write any new code and i made sure i didn't explore any new ideas in samba) and yet the other samba team members were not bound by the same restrictions.

so i believe i am entitled to be pissed off with open source - or more explicitly, how open source - and myself - has been exploited and violated.

to the extent that if i do develop and release a major project in the future with the potential to be attractive and useful to thousands or hundreds of thousands of people, it will be under a license that is otherwise unrestricted except to ban certain companies, such as VA/Linux, HP, Redhat, Caldera etc. from using or distributing it.

such a license would pretty much leave Debian and other truly non-commercially-controlled distributions as the means of distribution.

having a great time doing python coding. the framework that started off over three years ago as very very simple functions to output HTML tables from mysqldb has rapidly evolved three legs and two heads.

there's a company that did a commercially similar venture to the ultimate approach that i want to take: they're locked into a one year exclusivity deal on the code they supplied to their customer; the company since has gone tits up and the new owner wants hundreds of thousands of dollars for its suppliers to be able to break out of the exclusivity deal. reeally smaart lawyers...

in a few months that code will be utterly utterly worthless to that company, and open source will get a _really_ radical python/sql/html framework for application development that will, in my opinion, stuff zope, by comparison, somewhere into the C19th.

there's also ian's colorstudy "sqlobject" now on sf.net which is one small part of the whole python/html/sql thing.

the framework that _i_ am developing is also around - in sf.net/projects/custom - of a sorts. it's raw, it's rough, it works, it's not pretty, it's compact, it helps _me_ do applications.

i'm on to my 7th python / sql / html project (custom was the 6th). i'm not in the slightest bit interested in releasing this new project as open source at this stage: i have absolutely zero incentive to do so. i'm very happy to be able to say both parts of that sentence. open source can go stuff itself up its own arse: it's made me absolutely zero significant amounts of money so far. time to play a different game: time to change the rules.

maybe later i'll change my mind. but right now, having had my good nature taken advantage of so many times i figure, like in the nursery story where the chicken asks for help and gets none until it's time to eat, open source can go fuck itself.

... except in the story the chicken didn't swear, it was a very _polite_ chicken. maybe it was a duck. hell, i don't care: it was 25 years ago. for god's sake! shoot the damn chicken, feed it to the fox!

2 inches of snow in an hour, out of the blue (literally) and it totalled traffic conditions in watford.

m25 jammed so all traffic started getting off junctions 17 through 20 directly into watford; traffic not aware of this attempts to _also_ go through watford to get _on_ the m25.

four hours later after waiting a couple of hours to go home: wide empty and icy road outside the industrial estate. i've never tried doing hand brake turns before. it's fun!

[kids! don't steal your parent's car to do hand brake turns, they might get a bit upset if you prang the family runabout]

29 Jan 2003 (updated 29 Jan 2003 at 21:08 UTC) »

the "Custom" project i'm working on auto-generates virtually all of its HTML.

i'm looking at a way to separate the HTML formatting from the code itself, and found that there's a python program to do exactly that - with templates (apt-get install python-htmltmpl).

so i naively thought it best to start by outputting some of the auto-generated html code from a couple of example pages, and to go from there.

unfortunately, the functions used to generate tables, forms etc. are vastly more compact than the squiggly looking html output they create.

i've gone sufficiently far down the function route - display_form(), display_table(), display_header(), display_footer() etc. that it's going to be impossible to do entire HTML pages.

so i'm going to do specific templates: a display_form.tmpl that display_form() uses; a display_table.tmpl that display_table uses, etc.

and hope like hell it all hangs together.


if you like SIMD, take a look at aspex.net - theirs is a parallel processor on a chip with - get this: four thousand and ninety six processors.

those processors run at 250 Mhz, can do two-bit arithmetic that can be cascaded to emulate N-bit arithmetic of ANY length and i mean literally ANY length.

we're talking 4096 x 250 million bit-operations per second, here. that's one terabit-ops per second, out of a $USD 150 device.

... but it takes SIMD to the extreme: you REALLY have to think about your algorithm designs in a different light.

using the c++ template array classes really helps out there, as does making extensive use of python's functional programming operatives such as map and filter.


2 Jan 2003 (updated 2 Jan 2003 at 14:35 UTC) »

storms blow up around provocative articles. sometimes this is good - if handled carefully.

i had a quite lengthy and, for me, very informative, discussion with one of the people who responded quite vocally to the articles i wrote.

very interesting to see another article about open source - the title at least shows that i have made people think.

and for those people who are already willing to learn, well they don't need to read my articles, do they :)

and for those people who aren't willing to learn: if i made them shout out loud, stop or think, then that's more than i hoped for.

regarding the other article itself: chaos (evolution style) by numbers achieves more and is more self-protective than any other system. debian works in spite of its mailing lists...

Thief of Time

Interesting to note that even Terry Pratchett realises that there are people out there who think that calm and stability are always a good thing...

Custom. decided finally, after six months, to release this as an open source project. reason: am looking for a user-base and contractors and employees.

a quick search on sourceforge shows ZERO python-based e-commerce web systems (with any code). there is gnucash, but it's perl. there's compiere, which is java etc. there are plenty of php systems, and php totally sucks.

python is more than up to the job: my guess is that it's just catch 22.

well, here goes nothing.

yeubou, it's true! have you read "moving mars" or the ender's game series? sometimes i'm a bit like charles, in "moving mars", after he gets overloaded by the QL thinker, especially after two to three weeks of extreme concentration on one topic.

i definitely don't have the hypothetical anton's key thing described in ender's shadow (thank god). ... yes, if you've read the book that _was_ a hidden joke :)

i sometimes wonder if there was ever any significant brain-damage when i was born, because, hell, i just can't seem to focus on things that i don't consider to be important enough, and it drives people UP the wall.

[hey, btw, can you tell what new big words i learned last week?]

p.s. have you heard of transcendental meditation? it's a mind-control thing (starting with your own...)

p.p.s. that was another joke (for those people who think nothing that comes out of my mouth is funny or interesting)


well i am very pleased by the reactions to the articles i wrote: the responses have fallen into two categories. "i understand / i think this person is off their head" and "i think he's right / i think he's totally wrong".

very interesting to note that the split between these two categories is roughly equal, with there even being people in the "off his head" and "right" brackets...

the provoking nature of one of the articles was intended to bring out issues i consider to be important to the fore. one thing that hasn't really been discussed or mentioned, with people focussing instead on personal attacks, is the issue of perceived arrogance.

intelligent people need space in which to work with other intelligent people, and academia helps in this respect in some ways (and hinders in others). the level of expertise required for enormous projects is immense, and to have potentially hundreds of newbies and inexperienced people trying to help out is, in most cases, a complete waste of both the experts and the novices' time.

i wrote some new articles as background preparation for things to come. intended to be thought-provoking, also highly summarised (despite appearances!) because i tend to think in super-dense informational concepts and run the risk of expanding them and expanding them into too much details.

i find that the best way for people to get information out of me is to ask questions until they ask the right questions and eventually get the information they want, in the level of detail they want and can deal with.

too often the case is that people either dismiss the high-level conclusions i come to as being way too far-fetched: by that time they're not even prepared to listen or if they have questions, they don't want answers, only to demonstrate their own weaknesses.

usually to their detriment: information has a way of never being lost.

sadly the articles attracted some people i've never heard of who wanted to put their own reputations on the line by making personal comments about me.

i welcome any open minded and useful contributions these people have to make.

170 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!