There were a number of annoyances in movabletype, in particular it was too slow to post and getting slower and slower each time I added posts. Along with that I absolutely could not stand the way it managed templates and how totally cumbersome it was to change them.

Even more then that though, it's that I'm working on concepts for how you manage personal information that don't fit well with the rigid model that Movabletype uses. I believed there was a simpler and better way to build this type of system. Syncato illustrates a tiny amount of what I'm thinking.

The final reason is that I'm working on another project using Berkeley DB XML and I needed to put it through its paces in a real application. Note this is Berkeley DB XML, not Berkeley DB, they're related of course but not the same thing.

I don't think webware is CPU bound, it's simply the intrepreter lock is something to be aware of. Running two instances may or may not work depending on the application. Of course if you have an app that span can machines it will work fine.

So are you the same Geoff Cadian I knew in a former life?

Given the reason this spam is being submitted I'm still inclined to believe blacklists will be pretty effective. They shouldn't blacklist on connecting IPs, they should blacklist on the URLs included in the posts. The whole goal of the spam is to boost search engine rankings so the URLs have to be real to be effective.

The question of ownership is very different from the question of DRM. DRM is about control, and is an attempt by the media companies to control something that they shouldn't. When you buy a CD you may not own the music on the disc, but you do own the disc it self. You can be comfortable with the fact that you will be able to play that disc going forward because there's nothing to stand in the way of that (ignoring damage or theft). With DRM you're at the mercy of a central body that controls the future usefulness of the item that you're supposedly purchasing. With DRM involved you don't control any aspect of the equation. DRM is simply punishment to the purchaser and it's foolish that we have to endure it, especially given its complete ineffectiveness to task.

I certainly know how to circumvent the DRM on ITMS files, why should I have to? I paid for them, supposedly they're mine, but I lost control of them already with an OS upgrade. What's going to happen in ten years? What happens if Apple goes out of business? What happens if they get acquired? Basically to be assured of future proofing your purchase you have to jump through hoops and suffer an even further reduction in quality to strip something that shouldn't be there in the first place. This is crazy when what you're trying to compete with is free.

I've never been a big user of P2P networks, I've maybe downloaded ~100 songs and tried to replace most of those with legal copies. I own over 2,500 CDs and around 1,000 records. I now do all my listening from a computer and I want all my music in digital form. Given that, buying CDs is no longer safe and I really don't want any more CDs anyway. I want to be able to buy digital music at reasonable prices that I can be assured of being able to play going forward into the future without jumping through a bunch of artificially imposed hoops. That's not too much to ask.

Actually, I wasn't aware of that. Looks interesting, I'll check it out.

Syncato and the semantic web is something I've seen a number of people mention. Kind of funny since I'm anything but a fan of the semantic web concept simply because I believe it unrealistic in the real world. My goals with Syncato are much smaller, but if it makes some progress toward a more "semantic web", then I guess I won't complain. :-)

Native XML script from Bea is interesting as is the XPath support in Anobind (a little limited right now). The closest thing I've seen to what I'm looking for is still the XQuery implementation in Sherlock. I wish Apple would open that as a general framework even though I don't particularly care for XQuery it self.

Just to be clear, I'm not at all proposing that Python should actually be extended in this way. Python just happens to be the language that I currently use the most and it's just one in a long string of such languages. So using Python was just to be illustrative.

My interest in bringing this up is to give people something to think about when it comes to working with XML, or rather about how it could be done. I tend to take a different view of XML then most people. In particular I'm exploring concepts around using well-formed XML fragments as a data model that extends from persistence through to the user interface. Current languages are inadequate to do this as easily as it could be done. This is a concept that might be controversial to many developers, but it's also one I feel has an interesting future. Syncato is my first attempt at implementing the bare basics of this concept.

As for how you would query a relational database using built in XPath support in a language, you wouldn't. You would hand the query to the database and have it execute it just as you do now. The difference would be that you would ask the database to give you XML in return. You would then pick apart that XML using XPath built into the language.

Great thanks! Looks good with Movable type.

Anyone have a system other then movable type that can send trackback pings?

Syncato is XML from top to bottom and putting it on top of a relational database is a massive impedance mismatch. In other words it's not going to happen. It kinda defeats the whole point of what I'm trying to show. My specialization is native XML databases and Syncato was designed from the outset to showcase the possibilities of deep XML integration. I'm trying to show how leveraging a semi-structured data model based on XML provides maximum flexibility and results in a system of far greater power then traditional systems when you have the right application. Relational databases have schemas, and for the type of system I want schemas are not acceptable. Syncato plays fast and loose with its data structures and that's exactly the way I want it. It's a system for personal data management, not a corporate accounting system.

As for the URLs, Syncato does far more with its URLs then Movabletype does, it even does far more then systems like Bloxsom. I'm not really satisfied with the URL scheme and I've changed it a couple times already, but it will never be as clean as Movabletype simply because Movabletype just generates static HTML.

A lot of the URL scheme is driven by the REST architecture of the system. I toyed with the idea of using the bloxsom technique of using a virtual extension to change the rendering, but it seemed too limiting. I may change my mind about that again. As for the /WK/blog part that can be changed.

Well, the XML:DB API got a triple wammy of problems. It was intended to be database agnostic, language agnostic and also Java like. The DOM set the precedent, we definitely should have known better. :-)

Just removing the language agnostic bit would have helped a ton. At least it would have done away with the need to be expressible in IDL. I will definitely avoid ever trying to define another language agnostic API. It's just a bad idea that leads to a crappy result for all languages. Oh well.

Yeah it snows in the Mountains and the northern part of the state. It's kind of interesting because you get a suntan in the morning then drive for a couple hours and go skiing. It's too early for snow right now though. I've seen it snow in Phoenix maybe twice and I've lived here a long time. Never seen snow that stuck for more then a second.

If I do end up getting a camera you can count on some photo blogging features showing up in Syncato. No idea what, but there will be something, I'm way too lazy to manually manage pictures.

Boy do I feel your pain about classical CDs and the CDDB. I have probably around 1,200 classical CDs that are waiting to be ripped but aren't getting done because I hate fixing the meta-data so much.

I actually probably wouldn't try to use the plist file to change the meta-data. When you change it through the UI it changes both the ID3 tags in the file and the database. If you edit the plist file you'll have to reimport it and as far as I know the files won't get updated. This is why it's slow when you make a lot of meta-data changes through the UI.

BTW, if you select multiple tracks that you want to change to and choose get info you can change them all at once.

Yeah, I'm being a bit unfair, it was late. However, I still suspect they could have come up with a easier to use format that would have been just as fast to parse, especially since they control the parser. Of course, it could also have been a lot worse, they could have used a binary format so I shouldn't really complain. Anyway, this is a done deal so it doesn't matter now. We're stuck with the format.

Now about the XML support in Cocoa, that can be complained about. Or maybe not, at least in Panther there is actually some support without dropping down to C. It's still nothing more then the absolute minimum. This actually really worries me, I often wonder if Apple isn't missing the boat here. I believe XML is far more important then they understand. But then you look at things like Keynote and Sherlock and there is some hope. That just needs to extend into the core OS. Apple has the technology to have really good XML support, it's just locked up in the wrong places. The only thing that's promising is that a lot of cool features have started out that way and did make it into the core platform. I was just really, really hoping for much better support in Panther.

Well that was a little tongue in cheek trying to poke fun at people who are starting to have to deal with winter weather. :-) Phoenix is great and this is the time of year to be here too. Just don't come in August, that's when people everywhere else can poke fun at us.

I am always drawn to places that get a lot of rain just because they're so green. I think it's really beautiful. In Phoenix any green you see is artificial. I don't know that I'd really want to live somewhere else though, just visiting is probably good enough.

I'm well aware that you can use Applescript to do this. See the other post I referenced for a program that does exactly that. That wasn't the point of my post. My point was about how you deal with inadequate XML formats, iTunes was just an example I used because I had been complaining about the Apple plist format.

BTW, if you want to transfer your entire iTunes library for some purpose Apple Events are not faster at all, in fact they're quite a bit slower. That XML I was working with was 40MB representing nearly 30,000 tracks. The specific query would probably be faster, but again my point wasn't really querying iTunes it was about querying the iTunes XML file format. This is of interest if for some reason you wanted to use that information in a context where iTunes it self is not available.

I still maintain that most people can't tell the difference. :-) I used to be an audiophile and had a pretty nice setup, SACD, really nice turntable, electrostatic speakers, whatever. I used to sit people down and ask them what they heard. Invariably it was basically nothing. They could barely hear left right separation, let alone any kind of tonal quality or imaging depth even when I told them what to listen for. The one thing that they would consistently hear and be surprised by was a centered vocal that floated solidly between the speakers.

I once played Kind of Blue for my ex-girlfriend. Once on the original issue CD which sounds horrible and once as an SACD which sounds vastly superior, she thought the regular CD sounded better because it was louder. No, I assure you most people can not tell the difference. If you can hear the difference then you're more of an audiophile then you think. In reality though even if someone is capable of hearing the difference, the difference just doesn't matter in most environments where people listen to music and they won't know any better. This is of course anything but scientific data, but I've seen it enough that I'm pretty convinced that people just don't listen.

BTW, if you're listening to Saint Saens you're already well beyond what most people listen to. Also there is definitely going to be more of a noticeable difference in music that actually has dynamic range like classical and jazz especially with something like a pipe organ. But ... "most" people don't listen to music that has dynamic range, especially not pipe organs. You're also comparing two completely different recordings, with different performers, different halls and different recording engineers. There are massive differences just between CDs given all those variables and some engineers are much, much better then others. Even mastering makes a huge difference, this is why the original CD of kind of blue is so much worse then the SACD. Both were made from the same tapes, but the new mastering is far superior.

Oh yeah, my own experience is that I can absolutely tell the difference between a CD and any compressed file, but without really good equipment to play it on it doesn't matter that much to me either. The music is still great regardless.

Yep, bought a 10D. Really didn't want to, but I've developed a very strong interest in sports photography and the Rebel has too many limitations for that. I was realllly hoping Canon would update the 10D at PMA this week as the current model is not really that much of a step up from the Rebel. Didn't happen and the local shop actually had a 10D on the shelf so I bought it. Even though image quality wise it's almost the same as the Rebel, it's the features and build quality that makes it worth the difference. It's still not a great sports camera, but the 1D MkII was just announced and won't be available for several more months. The price of $4500 for the 1D is also a little more then I can justify. If I keep shooting sports I suspect I'll have a 1DMkII within six months.

Steven, the D70 looks to be an excellent camera. Knowing what I know now and having to make a choice between that and the Rebel, assuming both were available and I had no prior investment in lenses I would probably choose the D70. Just remember, you're buying into a system not just a camera. The camera's cheap compared to everything else. Well it is if you get bitten by the sports photography bug like I did. Fast lenses are expensive in both systems. When I started thinking about getting a different camera for sports I seriously considered switching to Nikon and getting a D2H, but even though I've only been at this for less then two months, I'm already pretty heavily invested in Canon gear. My dream camera is now the 1DmkII, I absolutely love shooting sports.

This gym usually works out around F1.8 1/250 at ISO 1600. Note, these shots were with a Canon 10D not a Digital Rebel. It's certainly possible to do indoor sports with a Rebel, but the camera makes things harder then they should be. If you're lucky enough to shoot in a gym where you can get away with ISO 400 then you can use sports mode and it works fine. Otherwise you end up fighting with the camera over AI Servo focus. I shot several grade school basketball games at my neice's school and was able to barely use ISO 400. My shots came out a little underexposed, but nothing that couldn't be recovered with post processing.

Really though, if you want to do indoor sports, the most important thing is to have a fast lens. I highly recommend the 85mm F1.8 prime for basketball. It's not too terribly expensive and it makes a huge difference compared even to a 70-200 F2.8L zoom. Something like the kit lens from the Rebel doesn't even begin to cut it. Of course, the other option is to add light with flash, but I've been avoiding that so far.

I have a 50mm 1.8 too and it's a good lens. It's not super fast focusing, but you can work with it if it's all you have. The biggest thing is that you have to have decent shutter speed to stop the action and in sports mode on the Rebel that may not be possible at ISO 400. Especially indoors. I used the 50 1.8 a lot at some 6th grade basketball games and the gym was just barely bright enough to use sports mode on the Rebel. If you want to do indoor sports the 70-200mm f4L won't be fast enough. The 85mm focuses faster then the 50mm since it's a USM lens, but if your problem is lack of shutter speed and you're already shooting the 50 wide open then it won't help any.

If your problem is shutter speed, you need to get out of sports mode and bump up the ISO. Try aperture priority mode at f1.8 and see what it sets shutter speed to then increase ISO to see if you can at least get 1/250 s. Also be aware at F1.8 depth of field is really thin so there's no room for error in focusing. Your other option is to resort to using flash to stop the action, though on camera flash looks really unnatural.

Speed skating seems tough. The skaters are in constant motion so you can't use timing to capture peaks in the action like you can do for basketball. I haven't shot anything like that yet, but here's a couple ideas.

Shoot at ISO 1600 and post process with Neat Image or Noise Ninja. That basketball pic is ISO 1600 and was run through Noise Ninja. Of course at web resolution you can't see the noise anyway, but I wanted to make a bigger print. If you shoot ISO 1600 just make sure you DO NOT underexpose the picture. That makes the noise much worse. The 10D and Digital Rebel are actually quite good at ISO 1600 as long you get a decent exposure.

Shoot from a position where the skaters are coming straight toward you. In that case you won't need as high of a shutter speed to stop the motion. Just remember at F1.8 depth of field is very shallow and the 50mm 1.8 might not focus fast enough to keep up with the skaters (assuming you can get the Rebel into AI Servo anyway). You might try manually focusing on a particular spot on the track and then timing the shot as the skater hits that point. Don't forget about shutter lag though, the Rebel is quick compared to a digicam, but it's not instant. There's a noticeable difference in shutter lag between the 10D and Rebel and the 10D is still slow compared to a 1D or Nikon D2H.

Shoot from further away. Might be tough at 50mm as the subject's probably already pretty small on the frame, but the further you are from the subject the lower the shutter speed that's required to stop the motion.

If you have any luck posts some pics.

Ian, iPhoto stores all the pictures in Pictures/iPhoto Library.

As a followup on this, iPhoto is now exhibiting another problem. Running it with a library containing about 20 pics results in a spinning beach ball at startup and it never recovers. I'd just been using it to print pictures. This version of iPhoto is pretty much worthless.