Notes and Jottings

“A look”:http://www.gush-shalom.org/thewall/ at the proposed location of the “wall” to separate Palestinians from Israelis. It shows some reasons why Palestinians might be grieved at the current state of affairs in terms of encroachment onto “their” land. We need a fair and equitable agreement, and an end to the jihad. A spiral into more violence seems quite possible though, given that neither side trusts the other, and also neither side seems to give any reason for trust.

… but if you did, you might find gems like this:

bq. You miss the point, XYZ is not an investment it is a cult and you are not yet a believer. Come join the faithful. Next year instead of an annual report XYZ will send shareholders a premixed glass of lemonade, shareholders will elect to join the kingdom of the faithful.

Amusing

[Scott Hanselman](http://www.hanselman.com/blog/) has been blogging about the Microsoft PDC and all of the upcoming stuff in Longhorn. Some items: the [first look at XAML](http://www.hanselman.com/blog/PermaLink.aspx?guid=b1853566-1d7f-429f-b926-5af79111f721) makes it sound a bit like a Glade XML UI description in GNOME. A [code example](http://www.hanselman.com/blog/PermaLink.aspx?guid=a3e34445-1a08-4a4b-a199-5f4da8fcd43e) makes it sound more like XUL from the Mozilla project. [Eric Meyer](http://www.meyerweb.com/eric/thoughts/200310.html#t200310024) thinks so too…

*First concern:* I hope that trust issues are considered if people can inject this kind of code into “HTML mail”. Otherwise we could have another security mess on our hands. This is the same kind of XSS issue that we routinely have to deal with because of active scripting being allowed at any point in a HTML document.

*Second concern:* I also hope that this does not kill off all competition in development platforms. It can’t, can it? We heard the signs of death approaching when *.NET* was the hype of the day. Diversity can also survive this. What Microsoft is doing sounds like good stuff, but nothing completely original as far as I have heard so far. I expect to see better-than-average implementations and excellent wizards though
What else? Avalon… a vector-based API for rendering. Again, not “new” as an idea (Nautilus has vector-based views), the difference is in the promotion of this as a “top-level API”, which is not really done with the SVG libraries in GNOME. I think. And the description of the creation of XAML code directly from _Adobe After Effects_ was interesting. I wish we could do that to create XUL code…

A “non-execute” flag for pages in memory. Isn’t this already a possibility on some architectures, just not on x86? Did I hear it was on the new AMD stuff (too tired to look that up)? This won’t stop all buffer overflows from being exploitable—there will still be a lot of writable executable pages—and if there aren’t, there could be code in memory which will do what you want, so just jump to there. But it’s nice to see this in there as a spur to get non-exec added to the Intel architecture. This isn’t a Microsoft thing really though, so come on Intel!

System.MessageBus—hmm there are aspects of this which sound a bit like D-Bus. But a lot more fully-fledged an idea I think. Sounds like Microsoft have a useful messaging architecture worked out in this.

Okay, I admit it. I’m impressed

This is the “deal maker”:http://www.oreillynet.com/pub/wlg/3930 for me. A new XML-editing mode for emacs which can validate against RelaxNG. Oh I really do have to get comfortable with emacs now! Years and years of vi and vim experience (and I’m very happy with them), but I’m not aware of any XML editing mode which is anywhere close to what emacs has had (in PSGML) and is about to have (in nXML). Are there any options for vim which are almost as good? Or is emacs the way I’m going to have to go? This doesn’t mean that I’d abandon vim altogether—though I will have to, for a while, just to help me achieve fluency in editing with emacs.

I decided this evening to go about writing a weblog server in Python that implemented the Atom API. As you do. Who knows how far I’ll get before I get distracted, don’t link to this because you think anything will ever come of it
But I have found that most of the time (in previous projects), my energy is spent either on low-level class hierarchy management, or high-level API parsing (should they be the other way around?)—and getting the program to the level where the two meet before I lose interest is a matter of pot-luck.

So tonight so far, I’ve been writing bootstrap code for the Berkeley database environment. Have to make sure it’s capable of being accessed from multiple threads of control, otherwise filesystem-based locking would need to be added. Free-threading introduces the possibility of using something like mod_python on Apache2. Also, local programs could join to the environment, for administration tasks which need to hit the databases directly. Transactions to keep everything in sync if I use reverse indexes. Aaarrgh, deadlock detection too! Deadlocks are the bane of Berkeley DB applications, if you haven’t been careful enough with your coding. It’s a bit easier these days with transactions—in the old days, you just had to make sure not to implement more than one database action simultaneously within a single thread of control (an open cursor could really screw everything up).

And after that, I haven’t gotten much further because I’m trying to decide how to divide up the databases, how to assign keys, what kind of user ownership model should be in place, etc. Seems to me like I have to work out fundamental questions like these before I go too far into the project. And it also seems that sitting down with a piece of paper and a pen is still one of the best ways of working this sort of thing out. Sorry computer, but I’ll leave you out for this.

“IT glitches”:http://www.unison.ie/irish_independent/stories.php3?ca=9&si=1067882&issue_id=9958 are causing a number of projects in the public sector to overrun spending by up to 300%. That is not surprising—but what is far less laudable is that IT systems purchased by different health boards do not interface (in at least one case).

Wouldn’t you think that after spending huge amounts of money on one of these systems, the data would be transferrable to another health board? This is indicative of how many people approach IT design—the bespoke implementation, so beloved during IT boom days, allowing the system to be exactly customised for a particular customer’s requirements, without considering the need for interoperability with other systems.

Thankfully, some people in the greater IT world are actively finding solutions to this, but I fear that many smaller (and larger) developers ignore the benefits of standardised schemata and common ontological definitions for describing problem domains. I read what I’ve just written, and I can see why people would say “What the f——”, but it is _so_ important, and yet something that people seem to come across by accident, if at all. Where is there a course in any Computer Science degree which describes approaches these types of problem? Don’t think it ever appeared in mine…

From comments to [an article on LWN.net](http://lwn.net/Articles/55317/) which espoused a belief as to what was wrong with current Linux, and what people should be aiming for. I haven’t read the article so I won’t comment on that, but this note was amusing:

> Since the superiority of micro over monolithic is so obvious, the only remaining detail is how to optimize the communication between the various components. Some have suggested XML, but I feel this is far too low level. If we put natural language processors in each kernel component, the communication could be done in a natural, and hence better, way. Now, this could be the source of some slight loss in speed, so the kernel components could have some embedded AI which would learn to use slang to speed things up a bit. This could be extended to user space, so that the NG Mozilla, instead of doing a malloc for another gigabyte of memory, could just do a system call which said (literally) “Gimme my usual”. And the kernel, knowing it came from NG Mozilla, would know what to do.

“British Pathe”:http://www.britishpathe.com has put 12 million still images from British Pathe’s library online; they were originally created between 1896 and 1970. As a tool for investigating possible historical images for personal use, it is an amazing resource. Prices are reasonable for non-commercial use—for example, £12.50 + VAT for a high-resolution still image. For commercial use, prices are higher—£25 for Powerpoint presentations and £500 for web sites, for a licence for 1 year. But I think it’s a fair bargain.

I had a dream last night about skiing—well about the night before the trip, and after having just arrived at the ski resort, eating some food at _Marco’s Bar_, and then heading off to go skiing, just no skiing itself.

I think Juan needs to tell us what’s going on At least I have one trip fully booked and ready to go for Christmas week! Yay, three cheers for a white Christmas…

Oh, we’re probably not going to Livigno this season, but check out “these reviews”:http://www.goski.com/rit/1livigno.htm. Is that a familiar name I see??

This was a bit of a film weekend for me, so some reviews:

* *Glengarry Glen Ross*: a film about personal and intra-office dealings betwen a number of people who work at a shady sales firm. Originally based on a play of the same name, the story relies on sharp dialogue and memorable characters. A simple story, but one which pulls it off well. There are two dialogues in particular which are extremely memorable, and excellently executed. Worth watching.
* *Ferris Bueller’s Day Off*: everybody remembers this one from the 80s. It has a slightly different feel these days. I wondered (aloud) about the hedonism it seems to promote—but then I was reminded (by others) of my lack of college attendance at times during college—so I’ll say no more… it’s still fun and energetic, have to cringe at some of the clothes.
* *Finding Nemo*: I thought I’d seen a bad review of this somewhere, but it’s actually just come out so maybe I didn’t. Anyway it was _much_ better than I expected, lots of laughs, the trailers do not do it justice. For all fans of _Monsters’ Inc_. No more spoilers from me though.
* *The Lord of the Rings: The Two Towers*: It’s almost the same as the book. Which is good. Faramir seems a bit meaner, which I didn’t like. I didn’t think Théodred was “ambushed”, didn’t he die at the Third Battle of the Isen? Éomer was banished? How did the Ents come so quickly when Treebeard called them to war? And he was far far too hasty for an Ent in any case! Elves at Helm’s Deep? And too much Arwen. I’m being picky. For the most part it was excellently done—Éowyn and Théoden are the best of the new characters introduced in this film. Gollum/Sméagol is just too annoying, but the character _was_ cleverly portrayed. +(Yes, I’ve seen it before, but I don’t think I said what I thought of it).+