@p


> I think it is hard to make it big and ugly while still organizing it sensibly.

TCP/IP could grow without AFAIK getting ugly because of its clear boundaries. Unix got somewhat uglier by growth by accretion. Browsers, maybe they started out ugly, idk.

> >one could actually see in the code that instead of ever taking anything out, they'd just add more

> Many such cases!

And I expect they all end up being impossible to even maintain.


> > because it was written for VAX

> Still runs on a VAX last I checked; I attended SCALE some years ago and the NetBSD delegation had brought a running VAX. Very fun.

I regret that I never actually got to play with one. The Math dept at Brown got one in about 1980 and I wanted to run some code in it (I'd gotten a manual and written something in Vax assembly), but Brown was the kind of place that would rather have it sit unused in a basement (what this seemed to be doing) than let someone use it who didn't have stampiti, even tho my father was a sort-of-big-deal Physics Prof (or maybe because: academic politics).
So within a few years I became a self-taught highly-paid OS coder at an elite Cambridge house. It never occured to me to go back and try to shove it in their faces, sort of wish I had.

> Being friends with him was more of a skill, but I chalked it up to normal Lisp programmer things.

I sort of knew two Lisp guys, rms and some guy at Symbolics. And probably the C-Interpreter guy, I bet. To be strange people among serious coders is an achievement.

> We had some disagreement about HTTP statuses (he was doing this backend data store with ACLs built from cons's of integers

Can't. Unsee.
and 0% of it dedicated to our application logic, his reasoning being that we could build whatever we wanted and he got to write a more complicated K/V store in Chicken Scheme; that was fashionable to do at the time: http://widgetsandshit.com/teddziuba/2009/06/startups-keep-it-in-your-pants.html ), and an hour in, I said "Okay, do what you want, I have to get this done." and the conversation then turned to how important it was that I never throw up my hands and accept any compromises to get the work done because everything must halt until we have unanimous agreement on what constitutes the Right Thing, which took another hour. Once you figure out how to be friends with him, he's a great guy, entertaining sense of humor, broad knowledge, but major stick up his ass about everything on earth. (I did manage enough self-awareness back then to recognize that I share this flaw.)

> ("Elizabeth", unless I'm thinking of someone else

I think you are.

> no one could work with him, is why Hurd became a historical footnote

His version is (I may be recalling incorrectly, treat this as half-remembered hearsay because it is half-remembered hearsay) that rms was somewhat uninterested in kernel development but very interested in micro-managing it.

> I'm coming over.

Ha, wait until the coughing stops, so that I'm certain I'm no longer contagious.