- From: Eric Johnson <ericj@nev.com>
- Date: Wed, 21 May 1997 10:31:53 +0100
- To: "W3C SGML WG" <w3c-sgml-wg@w3.org>
First of all, end tags are required. My work with electronic documents makes that very clear to me. I see from the traffic that I'm not alone in that view. Second, the argument to use short end tags to save space strikes me as being very much like the argument used that created the millenium problem, that is saving two columns on an 80-column card by not using the first two digits of the year. Someone way back surely said, "We'll never get to the year 2000 and anyway by that time, if it's a problem, it'll be resolved." Little did anyone realize. We aren't constrained by 80 columns any longer, but the problem remains as a vengeance. I predict that the use of short end tags will eventually come back to haunt us in a similar way. In that we can't anticipate all the uses tagged content will be put to in the future (especially permanent data) better to have the full information embedded in both the start and end tags than not. It may well turn out to be too expensive a problem to go back and fix later, to add full tags later as documents and data bases become increasingly complex. Either important content may in essence be lost or companies will be forced to ante up retrofits. My guess is that data will be lost at some not insignificant, bottom-line cost. Let's do it right the first time. Else: "Pay me now or pay me later," as the motor oil commercial says. While there does appear to be a genuine conflict between the requirements for the tagging of textual content and the incorporation of structured data (i.e., from data bases) in documents, it would seem that there must be ways of handling the problem that won't exacerbate the tagging of text by relaxing standards without very carefully working out the implications and not opting for current (say, immediate) and easier design solutions on the processing level. Eric
Received on Wednesday, 21 May 1997 10:31:26 UTC