April 28, 2003
Coddling bad software?
Should we bother trying to coddle bad ways of doing things simply because those are the only tools people possess? Or should we campaign to have them start using better tools and better practices?
I'm completely opposed to trying to jam RSS into HTML pages. It's certainly possible from a technical standpoint. So are a lot of things. But one has to ask why bother doing this?
Apparently the latest excuse is to allow people using HTML-only tools like Dreamweaver or Frontpage to create RSS. The idea, crazy as it sounds, is that these HTML folks will start jamming RSS-like markup into their HTML pages in hopes of it being readable by an aggregator. This is like soaking wood in gasoline to keep away the bugs and then trying to build a house out of it. Sure, it'll work but, in the words of Elmer Fudd, be wery, wery careful... One false move and fooom! The whole thing goes up. Insert other bad 'three little pigs' analogies here as well.
To try jamming this all into HTML pages raises the bar of complexity on all sides of the equation. Pages created this way will have to be much more precise. XML broaches no ill-formed documents, while HTML is rife with them. Likewise, existing RSS readers have their hands full just parsing bad encoding and stray text characters. To ask them to also become capable of tolerating all the abuses seen in HTML pages is no small request.
So why bother? To make it 'easier' for folks using pure HTML tools instead of template or data driven systems? How about recognizing the third angle, one a I see quite often neglected. Run a scraper against your own site. This way you can insert some trivial tags inside your HTML pages. You run your own scraper against it and immediately detect if there are any errors. The scraped data is put into a nice and simple XML file that the reader programs (and portals) can consume without leaping through flaming hoops.
There's a win all around here. The HTML tool folks only have to insert some very simple codes. Running the local scraper lets you see if you've made any tag mistakes right away. The scraped file is dumped to your own local drive so you save on CPU and bandwidth consumption. You readers get the sort HTML/RSS pair of URLs they're used to seeing. In short, you've bridged the gap between HTML tools and RSS without going down the path of XHTML insanity.
How is this not a better solution?
Bill: Is that Elmer Fudd or Chekov from Star Trek? :D
Posted by: Roger Benningfield on April 28, 2003 11:32 AMI believe his words would actually have been:
"Be vewwwy, vewwwy ca-ful."
It was "r"s he struggled with.
Boris and Natasha, of course would have used your version.
Posted by: Liz Lawley on April 28, 2003 05:27 PMHeh, that would indeed be Chekov. In searching for these sounds I was rather surprised to see a number of religious groups paraphrasing Elmer Fudd's 'be vewy, vewy, careful'. Odd indeed.
Rocky might have some advice.
Posted by: Bill Kearney on April 28, 2003 06:22 PMBill wrote:
>Run a scraper against your own site. This way you
>can insert some trivial tags inside your HTML
>pages. You run your own scraper against it and
>immediately detect if there are any errors.
Errrm, isn't this exactly what you're ranting against. XHTML-RSS is putting rss tags in XHTML to deliniate the items. You can then scrape it to something else.
Posted by: Ian Davis on April 29, 2003 07:24 AMIan, I think Bill means dropping (at the worst) some classes into HTML elements that you use to mark up your items.
Joe already proposed a using a namespace for XHTML Syndication here which is probably a better solution.
Posted by: Phil Wilson on April 30, 2003 06:39 AM
