Older blog entries for chalst (starting at number 149)

Past and Future
A week ago I met graydon and co in person: they're visiting Germany and neighbouring countries for a spell. Good to meet at last, less time to talk than I'd hoped, but we covered a fair amount of ground anyway. Most recurrent topic: what kind of frameworks are best for proving program correctness.

A week from now our household will be decamping to New Haven, CT for 12 months. Are there any advogatans in the area? ncm is about a 2 hour drive away, a drive I hope to make soon enough...

The month in LtU nodes
Flushing the backlog:

#3: Node #633: Dominus talk about HO Perl, with thanks to fxn;
#4: Node #645: Chemistry, Graph Tranformation and Programming Languages, about using graph transformations to represent the possible ways molecules can react;
#5: Node #663: Why do they program in C++?, about the mismatch between what PL theorists recommend and the enduring appeal of C++;
#6: Node #673: The Fortress Language Specification: another Guy Steele special, which in Sun's language for highly parallel programming, which reminds me a bit of Connection Machine LISP.
5 Apr 2005 (updated 5 Apr 2005 at 08:07 UTC) »
badvogato defies categorisation, but I spun the wheel and certified hir accordingly...

...I'm due two LtU nodes of the week...

...I'm in New Haven...

21 Mar 2005 (updated 21 Mar 2005 at 11:18 UTC) »
LtU node of the week, week #2
This week, I nominate Node #548: Journals and papers.

I cited this node in a posting, "Enormously Good News", to the slashdot story Wellcome Trust to Require Open-Access Publishing, saying about it There's been a lot of discussion at Lambda the Ultimate about the relationship of publishing and scientific organsiations like the ACM to the interests of the theoretically switched-on hacker community.

18 Mar 2005 (updated 20 Oct 2008 at 12:30 UTC) »

LtU node of the week
To try and get more regular about posting diary entries, I've dedided to adopt the semi-strict discipline of choosing to link to a Lambda the Ultimate node each week (either a new story or forum article, or something classic from the archives). I'll start this week with the very fresh.

Week 1: Node #587: The fate of reduce() in Python 3000, which has become a sort of forum for complaining about/defending Python's piss poor support of core FP style. Regular readers of this diary will probably have become aware that my attitude to the language moves around the spectrum from "distaste" through "irritation" and "outrage" to "cold hatred".

Thoughts on Recentlog
I'm reading the whole of recentlog, at thresh=2 much more often these days (my default used to be to just skim it at thresh=6). Very pleased to see ncm and raph posting again, welcome back! (And sorry, ncm, about the unwelcoming nature of my last post...).

A point: the reason I am reading recentlog more fully these days, is not having more time on my hands, but because it is faster to do, because fewer people are posting, and more rewarding, because the idiots have all moved to flashier ways of pulluting the blogosphere. Time, perhaps, for an advogato renaissance.

Pace raph, I don't think there is anything wrong with the trust metric, besides it's being meaningless-in-aggregate. Together with the underutilised diary rating mechanism, it really succeeds in keeping me stop reading crap on recentlog, with little personal effort, which is a very nontrivial accomplishment for an open community site. I'd say: experiment successful, but not quite in the way planned.

Today's responses on Recentlog
zhaoway on today's good laugh from wikipedia general equlibrium: I think economists don't tend to think hard enough about what they are taking for granted with the kinds of mathematical models they use (I'm a Neo-Ricardian, after all), but I can see why I might want to use assume I had at least as many traders [as] there are points on a real line when doing equilibrium modelling:

A trader is someone who performs actions in response to markets. Markets provide a sequence of pieces of information in real time. Hence an equilibium model ought not to behave differently when you assume the existence of all possible traders, who will number at least |(Time => Information) => Action|, which if |Time| >= Nat, and both |Information| and |Action| are >= 2, will number more than there are real numbers, or indeed continuous functions on the real numbers (both of the => funcation space formers should include at least some discontinuous functions under most reasonable topologies).
So, in short, I don't find the assumption ridiculous.

Read what robilad wrote about the purported complexity of open source licenses when compared the the reasonableness of Sun's offerings...

Google search for Siskind's Stalin
It takes a lot of qualifiers to get J. M. Siskind's ruthlessly optimising scheme compiler appear on the first Google hit for stalin, but "stalin -hitler -joseph -death -inurl:logs - USSR -Avidenko" does the trick...

Edit (2008-10-20): Fixed link to Stalin, above — I guess Advogato's html reformatting engine claimed another hyperlink...

16 Mar 2005 (updated 17 Mar 2005 at 08:10 UTC) »
ncm throws about Accusations...
Is ncm talking about what I think he's talking about, namely LtU node #531? Subnode numbers, please!

While I'm on Nathan's post, I don't think think the Common LISP community has been ideological for a good 10 years or so. Sentimental, maybe, or perhaps Nathan's mixed up CL with scheme?

Avriettea's Meta &c
avriettea oberved:

I can't help but think that when people read these diaries between all the military stuff and my ranting about the wikipedia and the gun pictures and stuff that I am freaking a few people out. When you say in your previous entry, you'd love to own the Barrett M99 rifle, and then you have a friend in OSD, and you hate a bunch of people, I guess somebody could get the wrong idea. I hope that's not the case. I'm a pretty boring guy, I promise.
And I couldn't help noticing that avriettea's diary appeared on my edition of recentlog at rating 2.2... Well, that's now fixed. I don't like guns, but I can appreciate gun nuttery when it's done right.

2nd CfP: Structures and Deductions
It's four weeks away from our ICALP satellite workshop's submission deadline (ie. 15th April). Attempts to make waves:

  1. LtU node #582: The domain is proof theory, the theme is "Eliminate Bureaucracy", the LtU angle is that if we can eliminate bureaucracy from proof theory, then we open the floodgates to applications of proof theory to computer science. That's the theory anyway..;
  2. The SD05 wiki page at the consequently.org wiki; and
  3. an entry at my Slashdot journal.
I wonder if any of this will reach interested ears...

Two Fine Polemics

  1. The Poor Man on Christopher Hitchens (via Ted Barlow);
  2. Scott McLemee on James Atlas.

    What was it that happened Christopher Hitchens? BTW, he's touring with David Horowitz in June, London dinner tickets a snip at $???? (deposit $500), strictly first-come first-serve basis...

14 Mar 2005 (updated 15 Mar 2005 at 00:21 UTC) »
Recentlog
avriettea (shit fixing) and avriettea (discouraged): I'm really very impressed at the way you are going about changing things you don't like with Wikipedia. I think I'm going to have to apply your example with some of the lameness that irritates me. I half wished I cared enough about modern military technology to appreciate the work you've done with your edits. Good luck with "system shit" (=DB server?). Oh, and congrats with how the M$ interview went (news?).

While I'm thinking about Wikipedia, I worked on the Trust metric and Advogato articles and created the Raph Levien page. I'm guessing some recentlog readers might like to contribute...

Signs of a Sick Civilisation
lemonodor has got evidence that's hard to argue with.

1 Mar 2005 (updated 13 Mar 2005 at 12:29 UTC) »
Recentlog
avriettea wrote: Pardon the shotgun weblog entry.
It's been months since I've enjoyed reading an advogato diary entry so much...
Oops spelling correction...
Curry-Howard wiki
Greg Restall has started a Curry-Howard wiki called "Proof and Counterexample"; it seems to be some kind of consultation exercuse for his planned book. I am enormously happy to hear this, which I heard of via, where else?, Lambda the Ultimate.

An Answer, sort of
Thomas Ptacek pointed me in the direction of ent, which isn't quite the low resource estimator for gzip performanceI was after, but it's in the right sort of direction.

Fallibility
I've been making lots of assertions that turned out to be false in my last few diary entries. I'm feeling particularly raw about misidentifying the SYNTAX-CASE records implementation...

24 Feb 2005 (updated 25 Feb 2005 at 09:06 UTC) »
A Question
Is there a fast way of getting a reliable estimate as to how much compression to expect from gzip? By fast, certainly I mean much faster than gzip in terms of cpu cycles/swap space needed.

A portable-looking SYNTAX-CASE implementation...
...has been posted to a SRFI mailing list by Andre van Tonder. I think I will port this to librep.

Nonsense, as Neil Van Dyke pointed out to me. I linked to an implementation of records in SYNTAX-CASE, which I mistook for what I was after.... So, for the record, I don't know of a fairly portable SYNTAX-CASE implementation, by which I mean one based on a reasonably easy to integrate low-level macro facility.

Spread the word
Peter Seibel's Practical Common Lisp is-out will come out very soon, and can be preordered. Via xach, ... and correction via xach's other, me-correcting, diary.

22 Feb 2005 (updated 12 Sep 2008 at 16:28 UTC) »
Wonderful!
Scott McLemee, who is essayist at large for Inside Higher Education, writes very, very well. There's a faint echo of Christopher Hitchens' in his writing, and the same non-conformist involvement with Marxist doctrine, but from across a cultural gulf (the nature of their respective educations could not be more different; [Bio #1] [Bio #2]), and without the unnaturalness of Hitchens' recent politics. Some samples:

NotSoRecentlog
zhaoway seems to be surprised I read his diary entry: actually zhaoway's diary entires is one of the half-dozen names who get me to (semi- )regularly read recentlog.

Postscript: While I'm on the subject of Hitchens (again): When the glitter fades... is worth reading.

140 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!