Getting it right

Big sighs all round. On xml-dev list this morning, there was an announce of an NVDL paper. here if you want to look. Source from MURATA Makoto and Hisashi Miyashita. The former is very much a known (good) quantity in the (small) NVDL world. He followed this up with a post responding to some of the misunderstandings on NVDL on the list.

The paper is all about using NVDL with atom (which I do, for my blog). I read it and tried to better validate my posts, then the integrated feed. Sheesh. What a mess. Out of about 200+ posts, there must have been near 50 that were invalid, when properly validated using NVDL. I updated my copies of onvdl and jnvdl to dig a little deeper, then when I finally got a working script I realised I'd been very sloppy with XHTML. I've written so little html that I'd been very loose in defining 'right' or valid. The translations were along the lines of

• quotation instead of blockquote
• blockquote without a nested p
• i instead of em
• b instead of strong
• q instead of cite

It really is embarassing. Anyway. Now fully re-built. nvdl now playing a real part in building my feed and I've hopefully learned a little more about NVDL.

I'm still not happy with atom, dropping an odd div element into the xhtml content element. Messy, to say the least. Leave the darned wrapper empty and let a decent validator such as NVDL despatch to a schema that can cope.