The great rewrite of Aquarionics is underway. Epistula2 is under construction, and there will be no more diary updates until it’s in a state where I can put them into *that*.
Xiven has now fixed the error that I mentioned earlier today (love this funky weblog thing), and I’m now working on stopping Epistula’s pingBack implementation from treating an XML-RPC error as a fatal one. I think we need a couple of error codes for that, actually. One to complain of a malformed packet (Which /should/ be a fatal error for pB, since it means my packet generation is screwed) and other to complain of a client error. Currently Hixie’s front page responds with an XML-RPC error if I pingBack it, which causes Epistula to think it’s mucking things up and kill the script (My own feelings on this are that if you can’t pingBack to a page, it shouldn’t have an X-Pingback server header, but since it’s possible – and done – I have to work around it (My server responds with an error string (Not an XML-RPC error) if it finds a link that can’t be pingbacked (pingedback? Pingback’d? Damn’ed is the language English for words technological. (Where was I before the endless stream of brackets, can you remember?)))(I thank my director, my agent, and my text editor’s “Find opposing bracket” feature)).
I belive in degradation. That is, if you read this site in Mozilla 1.2, it’s perfect. In IE6 it’s slow, but readable, in every version of IE and Netscape down to v2 it’s readable (And below that it’s only because of the virtual hosting of the server. It’s even readable on the travesty of technology that is Pocket IE (PIE ships on all WinCE devices, it has the HTML rendering & CSS ability of IE3, the scripting of IE4, and the XML parsing of IE5. Dreadful doesn’t cover it)). It does this with *no* server or client side scripting. It degrades, and uses the features your browser can provide. The point is not “Woo! I’m $foo compliant!” because I don’t care. The point is that I want this to work in as many browsers as possible without either serving special versions of pages up, or using HTML while conforming to all the standards that I can.
I can’t do that with text/xml, text/xhtml or anything more than I’m using now. XHTML1 may come in the future (Still served as text/html, because of the above), but right now I’m sticking where I am.
I’m now going to start running, before the twin evils of the parenthesis police and the Markup Police shout at me.
UtiliKilts. For every occasion, except bungee jumpting.
I have to admit, I did a lot of PHP when I was coding Epistula…
First, my foot is bleeding. Both my feet are bleeding. In many ways, this is my own stupid fault, but since I walked half an hour to the bus-stop this morning, and my new socks (With penguins on them) now have unattractive dark brown spots on them also. I hope there arn’t any sharks following the penguins.
Second, Mark has been playing, and now his 404 pages do exactly what mine do, only – natually – in a cooler and obviously far better way. But, because of his decision to pass XHTML1.1 headers to Mozilla, and the fact his page isn’t well-formed XML at the moment (Didn’t close a
) I can’t link to the article, because I can’t read his weblog unless I load up IE, and IE on this laptop isn’t set up with proxies or anything. I find this amusing. The “Broken in Moz” thing, more than the “IE not set up” thing.
I’ve been reading Hixie‘s Perfect Weblog specs, and disagreeing. Well, obviously validating markup is good. But most of all is the fact that it specifically says that if a page isn’t well formed, it’s broken. An example? Dive Into Mark is currently broken in Moz. So is this page which hixie linked to in his Perfect Weblog speech.
Aquarionics validates as HTML4 Trans. It doesn’t validate as Strict because the search form isn’t right (which I will fix), and the Blogsnob code contains a border=0 attribute (which I can’t fix). It doesn’t validate as XHTML1 because I now have almost nine hundred articles in the archives, most of which were written with EditPlus, which capitalises HTML tags by default. I’m vaugely interested in what he would say about the markup here (I’m probably going to submit it, merely for educational value while I’m doing the generation code for Epistula), but I’m not going to follow the world into XHTML until I split off the old archives or something.
The article discusses how people will like your program more if it’s polite. Not just that it says “Please” “Thank you” and “Sorry”, but that it actually predicts what you want next, tells you want it’s doing, and will generally be /nice/.
It’s something we don’t see enough of, really. There are few examples of actually helpful programs out there (The first version of Afphrid was praised for it’s helpfullness. Not sure about this version, nobody’s mentioned it), a good example would be Sam & Mark‘s RSS Validator, which tells you clearly what’s wrong and where to fix it. If any program or software I write can be described as both Useful and Friendly, I’ll be happy.
I’ll be even happier if it’s described as Bug Free 🙂
ccooke – my housemate – has updated his diary. Life just got terribly, terribly complicated. Twice.
Simon wants a way of saying “This person has more information on this” in a link. Aquarius doesn’t think this needs XML, or a defined format at all. I agree. My solution to this is currently being worked into Epistula, and is this: Crossreferences.
You may have noticed (You may not have done) that my “Pingbacks” box has been renamed “Crossreference”, this is going to become more relivant as I start putting in my Referer tracking, categories, and linking systems, because all of them just get more entries in the Crossreference box (I’m currently using pingback internally for when I mention posts within posts). It needs no technology, just a place in the design for that kind of thing.