If you want the version number just grab it from the app page. There is no need to implement a complex system to get it. Just get it from the server.
Unfortunately, screen scraping does require a complex system, at least on the client end. Consider that client software first has to pull out this chunk of HTML:
Code:
<span class="author">by Googer (port)
- 1.66MB, 1.0
- 25/4/2011
</span>,
That's easy enough, but then it has to get the version number, which is apparently everything from the comma on the second line to the end of that line. Okay, a client could find that, but what happens when you want to tweak the layout a little? Maybe you decide to just put it all on to one line instead of three? Suddenly the client can't find the version number. And that's only the version number we're talking about; that doesn't even account for title, description, rating, etc., any of which could break at the slightest layout change.
The repo spec is there to present all the same data, but in a way that a client program can easily understand it, independently of how it's presented to a human user. Presumably all the data is stored on your server in some layout-independent way, then Perl scripts assemble it into the displayed pages. Implementing the repo spec would therefore require just one more Perl script that assembles all that data into JSON rather than HTML. When I designed the latest version of the repo spec, I specifically had your app site in mind, and tried to make it as easy as possible to implement; though it may look large and hairy, the vast majority of its fields are optional, and you don't have to implement them if you don't want to.
Of course, if someone wanted to write an updater client that scrapes your site, that would certainly be cool. But, with PNDstore, I set out specifically to make a client for the repo spec. Attempting to shoehorn in site-specific screen scraping just wouldn't be worthwhile; a client would need to be built from the ground up. And I already have more than enough things to do, so it won't be from me. But still, I hope you'll give the
repo spec another consideration, since there's a client already waiting to use it.
