Older blog entries for ncm (starting at number 191)

IlyaM: <rant> What's wrong with Xerces-C++ is that it is an enormous leaky sack of converted Java code. (As Xerces-Java, it probably was well-enough designed, by Java standards; it's easy to find far, far worse.) Of almost 19,000 Debian packages, only three actual programs depend, directly or indirectly, on it; one of them is a recipes database, another is a toy web proxy. The Apache Project, if scrupulous, would withdraw it and Xalan-C++ from circulation, and officially orphan them. That they haven't done so yet is to their discredit. I've had two employers who suffered under it until they managed to squirm free. How many thousands of developer-hours have been spent trying to make it work in various projects, and how many more ripping it out of places it shouldn't have been put? </rant>

It appears that lo, after all these years, Orkut has solved its spammer problem. The only reason I have had, lately, to log into Orkut was to clear the spam from the groups I owned. Now there seems to be no reason left.

I wonder if Ohloh.net will really go anywhere. I see that GNU binutils is listed as used by only two members, myself included; I don't see any way for that to change, the way the site works. Anyway, I just created a project for Advogato, and told it about the mod_virgule SVN tree.

federico: As I understand it, that dialog is left over from old gnome, and would only appear if you made the mistake of setting X options in a Gnome control panel that doesn't exist any more. The Right Thing is to check "use X settings" and "Do not show". I don't think most people would ever see it. (But yes, it stinks.)

I just posted an article, "Is BitTorrent Evil?" So, in the strict technical sense, is it?

8 Aug 2007 (updated 8 Aug 2007 at 19:58 UTC) »

I thought etbe's questions, "Two Questions for All Serious Free Software Contributors", deserved an article, so I posted one. I think quite a lot of what people write in their diary^Wblog entries here would be more useful as short articles. There's no better way to increase the incidence of useful articles than to post them.

dmarti: The only reason I can think of ever to throttle a CPU is if it's running a "screen saver". (Quote marks because they don't save anything on an LC display.) Then, of course, the best clock rate is zero. On a related note, I think it's irresponsible that (e.g.) Gnome doesn't default to "xset dpms 0 0 1800" or similar.

OK, I joined ohloh. (It must be getting hard to come up with a name that Google has nothing else to say about.) Ohloh will be more interesting when they learn to count better; maybe Google will buy them and fix that and make it faster. In the meantime it's kind of fun. My only serious gripe is that they conflate C++ and C code together as "C/C++"; "Java/C#" as one entry would make more sense. Maybe the best thing about it, so far, is the map: when you tell it where you live, it shows markers for other Free Software people nearby. It will get better as more people join. That would be fun to have in Advogato.

Today I deleted the Orkut group "Stupidity". It was the only one of my Orkut groups that achieved membership of more than 1000. (Curiously, the overwhelming majority were Pakistani nationals.) However, it utterly failed in its purpose, which was as a place to discuss Prof. Cipolla's theory of stupidity, with its enigmatic constant σ, the fraction of any given population, however selected, who can be counted upon to exhibit stupidity. (There may be some irony there: quite possibly the value of σ found in that pocket universe would have sharply contradicted the hypothesis of its constancy.) The theory, meanwhile, is in eclipse, as the professor's literary executors have worked diligently to scour what they claim are immature drafts of it from visibility on the 'net, effacing him from public memory in the process. It's not clear whether that is also ironic.

I noticed recently that dmarti's diary entries were marked with an interest level of "1", obscuring them from most readers' view. I added a "10" to the mix; I hope that's enough to restore him to visibility, but he may need others' help too.

I notice the OpenMoko has a 283 pixels/inch display, which should be of some interest to Raph, at least.

etbe: ... and you can pump heat out of a big solar water heater tank to produce a smaller amount of hotter water. If you can bury a big water tank, you can use Ted Taylor's scheme to save up cold from the winter by misting brine and draining the slush to the tank, to use as a heat sink in summer. You can pre-heat water going into your solar water heater with heat extracted by your heat pump while cooling your house. Heat pumps can be remarkably versatile once you get the plumbing right.

I've had a new insight about atmospheric carbon dioxide. It's not something you'll see published elsewhere, because it's nuanced, and public discussions of global warming don't tolerate nuance. Anyway: the great weakness in atmospheric simulations is that they have no way at all to model cloud cover. Nobody knows what the effect of anything will be on cloud cover, despite that cloud cover affects absolutely everything climatic. (Some people have pretty good evidence that cloud cover is affected by solar activity and even by cosmic ray flux.) This means that even the very best models can't really tell us anything about the climatic effects of CO2 under conditions where cloud cover is changing, despite that a different climate will certainly have different cloud cover.

Why, then, should we worry about CO2? First, it clearly has been a primary influence on runaway global warming so far, and that ought to be enough to freak out anybody. (Cloud cover shows no sign of jumping up to save us yet!) But even if cloud cover were to jump and save us (and, who knows, plunge us into an ice age?), CO2 is certainly a direct cause of coral bleaching (via dissolution in the ocean, forming carbonic acid) and the crash of reef ecologies, the basis of the ocean ecosystem. We take millions of tons of food from the ocean ecosystem every year. Any sane person should find that enough reason, alone, to move heaven and earth to cut CO2 releases. Uncertainty over the effect of CO2 on the climate only redoubles the urgency.

24 Jul 2007 (updated 28 Jul 2007 at 00:09 UTC) »

I'm just back from the C++ Standard meeting in Toronto. For the next ISO C++ standard, the Standard Library will have regular expressions, threads, and sockets, and the language will have lambda (i.e. code in brackets passed as an argument) and auto variables (i.e. the variable or argument declaration takes its type from the value used to initialize it). Equally interesting, but mainly to library authors, is concepts, which allows writing (even more!) powerful libraries that give actually-readable error messages at compile time if you misuse them. I say "even more" because with concepts the library can compile to different code if it's passed (e.g.) a literal string than if passed a pointer. The combination of regular expressions, lambda, and concepts means we can write yacc as a C++ library, and offer both better error messages and faster parsing than yacc can.

182 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!