Older blog entries for chalst (starting at number 82)

zhaoway: you are being rather mysterious. Don't worry about Erik, he only bites posters to comp.lang.lisp. In any case, isn't it better to get bitten early than suffer deeper bites later?

tk: Indeed you are right, it's not so difficult, and I'm not sure what Mike thought the problem was. Perhaps the point is that you can't parse Haskell/python using lex/yacc without the use of persistent state in the lexer. The right data structure to handle indentations is a linked list; you can then count the number of times you pop the list to see how many OUTDENTs to issue.

Terrorism and many eyes
Oliver Kamm writes about the idea of a futures market where one can bet on future terrorist attacks. He argues the plan is incompetent, because it tries to make predictions in the absence on public information. However, he says a similar idea would work, namely using market mechanisms to identify risk areas for terrorist attacks.

There's an interesting point of similiarity between this idea and the idea that open source projects tend, others things being equal, to be less buggy than closed projects. Both have the criticism that attackers can use the published information to find vulnerabilities, and the counter-argument is comparitively complex.

28 Jul 2003 (updated 28 Jul 2003 at 17:33 UTC) »
Indentation-based LISPs
zhaoway writes about his ideas for an alternative syntax for scheme. There's been some experimentation with Haskell/Python -style indentation based syntaxes for LISPs. Check out:
  • This USENET post of Paul Hernhout from 2000 outlining such an example.
  • This post of Peter Norvig, from the same thread, arguing for the naturality of indentation.
  • This post, also in the same thread, by Eric Naggum argues that indentation based syntax is a bad thing for LISP-like languages since it breaks the important similarity between internal representation and programmer representation that is so important for macro programming.
I also remember reading in comp.lang.scheme about someone who used such an approach in teaching scheme (training wheels for LISP syntax?), but I've lost the reference.

zhaoway also comments on the difficulty of getting lex/yacc to handle non C-like syntaxes, by which I guess he means indentation-based syntaxes. Yes, this is an acknowledged problem with lex/yacc: Mike Spivey was the first to tell me of the problem. It's not impossible to do this: lex does allow you to call arbitrary C code when lexing; but I think there is no clean way to do this. Maybe, as a challenge, some lex guru here can show a clean way of doing this.

Postscript: Politics on Advogato
kilmo sort-of reversed his earlier stance against politics on advogato stance due to irritation at ignorance, to post the first in a promised series on Israeli history. It's a nice start, I have some quibbles, but he hasn't yet touched upon the really interesting part of the regional history (and Daniel Bernstein's guest post at the Volokh conspiracy covers more of this history). I'd like to make a couple of points about politics on advogato:

  • Firstly, I don't want to run different weblogs, I like being part of the advogato community, and I want to post about politics. The idea that there is a tension between content on free software and other topics has led raph first to propose a system of sections, and later to create his own personal blog server. The first didn't catch on, the second I don't like since it means I have to read several different places to keep track of all the things that are going on with raph...
  • Secondly, just how bad for the free software are posts about politics? I'm guessing that the suspicion about political content here has been due to mglazer's abuse of this forum. But here I think the problem has been not with mglazer's views per se, but the offensive manner in which he frequently framed his views. While we have had problems with other political views put across offensively, I think we have had less problem with that than with offensively made views about developer and free software issues, and I think that diary ratings have pretty much solved the problem.
So I think there is a place for politics on advogato. Am I out of touch on this?
Politics #1: Andrew Giligan
It has been widely claimed by BBC haters that Andrew Giligan during the invasion of Iraq was a defeatist reporter, gleefully exaggerating setbacks for US/UK forces and failing to recognise accomplishments. Interesting to learn that he was a supporter of the invasion: see this article by Boris Johnson MP.

Politics #2: The Bush administrations *real* motivations for invading Iraq
John Marshall's Talking Point Memo provides an interesting discussion of the real motivation for the invasion of Iraq. I was convinced that the motive Josh gives (roughly, reengineering the Middle East to promote democracy, and so make the region safe and friendly to Israel and the US) was the real motive for the war since I read this interview in Foreign Policy. If this is right, then it is clear that the Bush Administration as a whole deceived the american public, despite having honorable motivations (Paul Wolfowitz being the decent exception, who did make this case).

Josh says the reason for the deception is the administration thought their actual reasons were too complex to make to the american public. One is reminded of Dean Acheson's infamous remark about the need to explain the case for going to war in Vietnam "in terms clearer than truth".

15 Jul 2003 (updated 30 Sep 2008 at 23:14 UTC) »

Larry Solum, a law blogger, has begun a series on intellectual property on the divide between academic legal theorists and practicing lawyers, the first post responding to arguments at Scrivener's error that, irritatingly, he doesn't link to. I think the series is important: by and large IP theorists are on the side of weak IP, and practicing IP lawyers are against it.

Postscript: Are two-layer UIs a good thing?
It's a common pattern in applications that they are given two interfaces, a deep interface that exposes as much functionality as possible, and is usually made available as a library, and a shallow interface that exposes only a subset of the functionality needed in most common user interactions, and which is available as a command-line argument interpreter, or as a GUI.

There are generally good reasons for adopting this pattern, but I wonder if it is overutilised through familiarity. Two thoughts:

  1. Why shouldn't we expose all the functionality in the user interface? Embedded scripting languages and their LISPish ancestors make this kind of thing easy, and you can even have dialog boxes in GUIs if you like.
  2. I'm also intrigued by what the Mozilla folk are doing with XUL. They don't see there as being *one* shallow interface to the Mozilla code base, but a whole family of them. Everyone can write their own browser-like application through XUL, so we have an application as domain specific language.
Obviously, flexibility doesn't make good UIs, but I wonder if we might not get UIs that better balance the needs of both ordinary and expert users if we go someway down this path?

Postscript #2: VIPS
The Veteran Intelligence Professionals for Sanity caused a bit of a stir back in April contractadicting claims of there being a consensus in the intelligence community saying Saddam Hussein's regime had an active nuclear weapon's development programme. Now they are calling for VP Cheney's resignation. Does anyone know anything about this group, other than the usual boilerplate one hears that they are a bunch of mostly retired CIA analysts spread across the USA?

Postscript #3
Unbelievable. (In case you think this is just common or garden stupidity, bear in mind this is the front page story of the UK magazine that styles itself the magazine of choice for the intelligent conservative).

Postscript #4
Felix Salmon linked to PS3 above in his article Pyramid schemes in the Spectator.

23 Jun 2003 (updated 23 Jun 2003 at 20:08 UTC) »

Mark Glaser has done a chart of the most important blogs according to impact on print media. The blogs are arranged from liberal to conservative, and from bloggish to journalistic. It's a nice overview, though it has makes some judgements I find suprising: how is the democrat Mickey Kaus (Kausfiles) supposed to be more conservative than AndrewSullivan? And is the Drudge Report still one of the six most influential blogs?

raph: glad you liked the link; I think the issue is important and I'm happy to hear that you are thinking about how to counteract the effect. I have to confess the text with my link was a piece of unsuccessful irony, because I found Harry Frankfurt's tone rather pompous, so the piece might be seen as a bit bullshitty itself. But I do recommend folks read the essay. Also Orwell's Politics and the English Language is strongly recommended: although bullshit isn't really his target here, the mentality that lies behind it is very much addressed

17 Jun 2003 (updated 30 Apr 2005 at 11:30 UTC) »
On bullshit, a lucid and humane exposition that succeeds in bringing discipline and intellectual rigour to this crucial, but underanalysed, category of discourse.

Postscript (30 April 2005)
Kieran Setiya talked about the clever rhetoric of Frankfurter's essay, clearly much cleverer than the failed irony of the above...

Stevey: I think that diary ratings might provide a better basis for weighting than ranking, since it provides some measure of the persons sensitivity to the advogato audience, but simply asking for a single Master to second an article is probably the simplest measure that would work. I'm mildly against having negative votes for articles: if the concern is emptying the queue there are less divisive ways of achieving it, eg. saying once five later submissions are accepted, the article expires.

What's the problem with posting comments to pending articles?

21 May 2003 (updated 21 May 2003 at 19:32 UTC) »

Aha! Even my jaded tastes found this new trend interesting...

14 May 2003 (updated 14 May 2003 at 19:22 UTC) »
ncm: My "Null Hypothesis" asserts only that fxn's general tendency is not to overcertify Master's, not that he always doesn't, by analogy with the role that null hypotheses play in statistics. I'd say, also, that if you are not a Master, you are certainly in the upper ranks of the Journeyer's: it's really a judgement call, so you are not a bad exception to the rule, if you are one at all...

cbbrowne: Welcome back!

Anti-Americanness in Europe
A good antidote to the idea that lots of people express that the European and American paths are now separating is Eric Alterman's article in The Nation USA Oui! Bush non!. The general gist I take from the article (and agree with) is that while the Bush administration has done a lot of harm to relations between the US and Europe in the short term, the harm done doesn't go very deep since it reflects the views of a relatively small elite in the USA, and there are enough reserves of goodwill to repair the damage, on both sides of the Atlantic, when the Bush administration is gone.

I don't suppose there is any chance that mitnick is genuine? Reflects rather badly on the advogato trust metric in any case...

30 Apr 2003 (updated 30 Apr 2003 at 16:47 UTC) »
graydon's law(?)
graydon writes: programming language features do well (all other things being equal) when they eliminate either distant or dynamic state and replace it with either close or lexical state.
I think this is true, and a smart observation that seems to explain quite a few trends over the past 30 years, though I disagree with graydon's explanation. My suggestion is that it becomes easier to make true assertions about such programs, both for humans and by compilers, when we make these substitutions.

W.r.t our ongoing and inconclusive argument about static vs. dynamic typing, I'd like to say that dynamic types make it harder for machines to make these assertions, and they do not discipline programmers to make these substitutions, compared to static types, but my objection to static types is that they achieve these goals at a high price in the complexity of program construction in cases that occur rather frequently.

Orwell unsainted
Some kind advogatan pointed to me an interesting story of disillusionment with Orwell. Unfortunately I've forgotten who that is and it is not on recentlog any more: please email, whoever it is. Postscript: Google finds it: it was wlach.

The Null Hypothesis for overcertification
It's been commented already that fxn has certified a lot of people, but I don't think it has been commented on just how good his certifications are. Since there is some concern here to the effect that there are too many undeserved master certifications, can I suggest the null hypothesis that a master cert is an overcert if fxn rates that person Journeyer or less? I have found a few people I think deserve Master that he certs Journeyer or less, but I haven't yet found any counterexamples to the converse.

73 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!