URLs
I know I have way too much free time on my hands. but I am again contemplating switching blogging engines.
as lovely as Blosxom is, I'm not entirely down with the file-format. it is admittedly very nicely simple: the first line is the title, the rest is output as is (in my case, garden-variety HTML, but other bloggers have implemented other markup languages [ATX][textile][wikiwordish].) but, unlike XML, it is not as easy to spot errors in the markup, particularly with the pre-theme flavor files. (and if you dig around my site for a while, you will see instances plagued by HTML errors such as unclosed <a>
tags and mismatched <span>
tags.)
so, because of my masochistic nature, I am trying to write my own blog engine based on XML and XSLT. I am a little apprehensive as to what sort of load this will cause to my webhost despite the light traffic, but there is always static rendering, I suppose. In reality, I'll just be writing a more dynamic version of my hand-kludged blog engine that I had been using previously, with more Perl and less XSLT.
I'm still stuck on parsing the requested URL. I have decided on the popular format of /category/subcategory/2004/02/07/slug
, although I'm still wondering about the extension. on one hand, slug.html
is kinda crufty looking. on the other hand, I like Blosxom's notion of flavors. still, the only other flavor that I can think of that would be worth supporting at this time is RSS, and while good for indices, I can't see the utility of serving up individual RSS entries. and this in itself leads to some parsing ugliness. if I stick to the flavor idea, then my RSS feeds will be specified like /category/subcategory/2004/02/07/index.rss
, and I realize that index is pretty crufty. but if I discard flavors entirely, /category/subcategory/2004/02/07/rss
would probably be how I should specify a feed, but I think it would cause a parsing nightmare. I think I would have to hard code it, but that also strikes me as inelegant.
so it's still a work in progress. I've spent a couple of days on the parsing routine already, using Blosxom's routine as a take off point. I've also been screwing around with regex's, but I imagine that would be processor intensive. so we'll see.
anyway, I found inspiration about non-crufty permalinks here:
- Cruft-free URLs in Movable Type by Mark Pilgrim
- Blosxom's two types of permalinks by Nelson Minar
- How to recognize a Weblog tool by its permalinks by Matthew Thomas
Article URLs week by Nathan Ashby-Kuhlman Principles Readability, Brevity, Cleanliness, Hierarchy, and Permanence Awful vs Respectable URLs Dates need to be arranged in year/month/day order URLs at some of the top news sites More URLs at stop news sites Recommendations Nouns and Verbs(interestingly, his description of nouns and verbs corresponding precisely to the OOP concepts of objects and methods. to rephrase his point, objects need to be bookmarkable, but methods don't, which makes a lot of sense and, is in fact, almost tautological.) </ul>