title:      James Clark is a Sainsbury
date:       2022-11-22
tags:       history  notes  phlog
identifier: 20221122T175630
---------------------------
I was thinking about XML this evening, watching a braindead bro on irc
talking shit about it and saying almost nothing over and over again. I
wanted to chime in, because the main inhabitants of this channel were kind
of annoyed, but instead it sent me on a little wander down memory lane.

James Clark wrote expat, the parser that's everywhere in GNU/Linux userland.
He also wrote XP and XT, which are two extremely un-Javalike Java libs, for
XML and XSLT respectively. He also wrote the GNU troff (groff) - his first
major project out of school. He donated it to the FSF when he was done.

I found that last part out the other day when I was messing with groff to
format something for gopher. From about 1999 until 2002-ish, I spent a lot
of time with those Java tools he made. I was surprised to find out about
groff.

On that same outing, I was remembering that he had lived in SE Asia (not
sure if he still does), like another fine hacker and all-around
over-achiever, technomancy (Phil Hagelberg). I recall sometime around the
turn of the century back in Camberwell, reading possibly the same bio page I
was looking at the other night. Probably it was a bit shorter. Back then I
saw him as an example of what I idealize, which is a person who goes where
their attention takes them and as a side effect supports the life they
choose. I had a sense of SE Asia being a lot cheaper than SE England to live
comfortably, and it was inspiring to think about a life like that. But then
I read on his Wikipedia page, which was NOT a thing back in the early
zeroes, that he is a member of the Sainsbury family, founders of a 6bn GBP
supermarket chain, of which they still own about a fifth. Tangentially (but
also interesting) they seem to have been selling down their stake in recent
years - 1/4 is now owned by Qatari interests. More sparking, more yaks.

So in some sense I feel a little cheated, because Charterhouse, and then
Oxford... this is not someone who has needed to work for food. Don't get me
wrong, I still admire his work, those parsers were excellent, well
engineered, deployed a lot of places, doing high throughput work. And I
admire the work he does, and work product he donates, toward ethical
software practices. But he was doing good works, not working for a living.

Maybe I'm wrong, or maybe he started out that way and eventually earned his
way, but that's the rich white dude lottery doing the work. I definitely won
that lottery myself, though not to that degree. I just got everything I
needed plus a lot of other enriching stuff, none of which was free. But I
went to state schools and my folks weren't big equity people. Not hand to
mouth, but not capitalist enough to get proper rich.

Anyway, software history turns out to be my history too, which can make hot
takes from some 22 yr old software goldfish tiresome. There's a lot to
dislike about XML[1], I barely go near the stuff any more if I can help it,
but when I think back to what came before - well, packed binary wire
protocols between banks, and occasional large corporations, on leased or
even privately owned lines. No way to do what they did with any degree of
privacy or reliability over public networks. Brittle, codependent, a risk
nobody wanted to take. XML provided a toolkit for rapidly prototyping and
evolving new interactions, without crashing all the time. If you wanted
validation, then it would cost you, and things got brittle again. But if you
were ok just throwing out docs or stanzas that didn't parse legally, and
doing something else with the bits you just didn't understand yet, you could
go at breakneck speed and make genuine progress.

And! You could trivially display it in a structured way, which is not
nothing.

Some of what followed was truly monstrous, but show me some dimension of
tech progress that doesn't desperately need Jeff Goldblum to question
someone's judgement.

James Clark wrote some dope code, is all I'm saying.