Aha.
In ASP.NET 1.1, you need to have a folder called aspnet_client\system_web\{.NET version} under a Web application's root in order for Javascript to work.
In ASP.NET 2.0, you don't. And in fact, on a server (like mine) where both versions are running side by side, having that folder causes Javascript to fail in some browsers on the ASP.NET 1.1 sites (like this one).
This means comments are working now.
But I'm still going to install Community Server, though I probably will keep Das Blog now. (For the record, I always thought it was a configuration error, not Das Blog's fault.)
I finally
bit the bullet and downloaded the Visual Studio 2005 CD images from Microsoft, and installed the latest runtime on my Web server.
Only one site broke: Hired Wrist, my dad's site, which I just now fixed. That's not bad. Usually upgrading hoses everything.
Hired Wrist broke (gracefully, I should point out; only the graphic headers were affected) because the released version of ASP.NET 2.0 handles page names slightly differently, which caused my resource-based graphics handling to fail. Resources, apparently, are now case-sensitive. Oops.
I've just spent the past four and a half hours trying, and failing, to get Microsoft SharePoint installed and running.
I think the .NET 2.0 Beta runtime on my main server is screwing things up. I think this because, for example, other people have gotten SharePoint running without a problem, and my Das Blog difficulties only seem to affect this server. (I got Das Blog running on a laptop—which doesn't have .NET 2.0 on it—just fine.)
Why doesn't stuff just work?
I'm all ready to start testing two open-source prouducts that are built for .NET 2.0, which was released about two weeks ago. I can't yet because I don't have the final version of .NET 2.0 yet; I still have the final beta, and these open-source projects won't run on the beta.
My company subscribes to Microsoft Development Network, which gives us just about everything they sell, plus all the beta-test versions. They also have a site from which we can download anything we haven't received yet.
So today, when I finally have some time to play with the new stuff, their download site is down. And we haven't received the DVDs yet (they're due in about two weeks). So I can't do much of anything that I wanted to do today.
Fooey.
About every five years I learn something about my craft. This is an average; the last seismic shift happened in 2002, but the one before it happened in 1995.
It's happening again. This time, I'm learning how my craft gets in the way of my business.
For the past three years (since the last time a two-by-four hit me) I've worked on the Inner Drive Extensible Architecture™, a comprehensive framework on which Inner Drive can build marketable applications. It's a masterpiece, in the way a fine, ornate table would demonstrate the competence of a 17th-century carpenter.
It turns out, I missed the advent of steam.
See, other people have already done it, and they're giving away their code. So over the past few weeks I have slowly come to realize that there is no point continuing this effort in the same way.
This demonstrates a constant, historic tension in business software: Build or Buy. Build something exactly the way you want, or buy something that's close enough.
Open source software, like Das Blog and DotNetNuke, makes this choice even more stark. Can one countenance spending 600 development hours on creating something that has half the features of something anyone can download for free?
More on this later, as I refactor the IDEA to extend, rather than duplicate, the mass of free stuff out there.
The Code Project has today publicized details about Sony's DRM CreepyWare that lets Sony know what CDs you're listening to. It also hides in the bowels of your Windows operating system and can't be un-installed without downloading a buggy patch from Sony.
I'm all in favor of protecting copyrights. But this is creepy, and more offensive than the Mickey Mouse Protection Act of 1998.
Update: The L.A. Times has the story now.
(Last post for now.)
Anne just sent me a link to a very cool application.
That is all.
Update/Clarification: Anne found this to aid her writing, not because there's any special news from our family. But thanks for asking.
Translation: I am going to replace Das Blog.
Yes, as much as I like it—and I do, despite my gripes—I've found something more in line with the way that I work: Community Server. I've downloaded both the current version and the new Beta, and as soon as I have time (tomorrow at the earliest, Saturday at the latest), I'll switch over.
Since this is a brand-new blog, I have little compunction about wiping out all the permalinks. (There are fewer than 10 at the moment.) We apologize for the inconvenience.
Which do people prefer: having all the entries on the front page, one line per entry (like they are now), or having the complete posts all in a column? Talking Points Memo is an example of the "flow" design.
(I also still don't have the comment problem resolved, so please email me with your thoughts.)
Thanks for your help.
I have to give up on converting dasBlog to .NET 2.0 for now. The ASP.NET model has
changed significantly from 1.1 to 2.0, breaking every single page in the application.
In other words, many of the techniques developers used for making pages do interesting
things in 1.1 simply don't work in 2.0, because the system doesn't work the same
way. I may have missed something simple—but I don't think so at the moment.
So I think all of the pages in the application need to be changed in order to make
it work. In fact, I think the whole application needs to change, to take advantage
of the new features in ASP.NET 2.0. I figure fixing the current pages will take
about 8-12 hours, which I can't spare until Thanksgiving. I'll just have to live
with a few annoyances for now. Redesigning the application isn't going to happen.
Tuesday I'll take a look at it in Visual Studio 2003, so I can at least see how
it works step-by-step, and possibly see if some of these annoyances will take a
lot of time to fix.
The glimmer of light in all this is that only the Web application needs to change. All of the supporting assemblies compile without any major problems. So instead of changing hundreds of classes, I think I only have to change a few dozen.