- From: David Ornstein <davido@apocalypse.org>
- Date: Mon, 29 Jan 1996 12:17:58 -0800
- To: bsobilof@inet.ed.gov (Blake Sobiloff), www-talk@w3.org
At 02:47 PM 1/29/96 -0500, Blake Sobiloff wrote: >At 1:12 PM 1/29/96, M. Hedlund wrote: >[Description of <INSERT> structure] > >>Advantages: >[...] >>* Doesn't stuff all possible variants into one document (avoids filesize >> bloat). > >I've heard a lot of people mention filesize bloat as a (potential) problem, >but I wonder how much of a problem it really is. At least with the files I >ride herd over, I doubt that many would more than double in size -- but >then again, I'm pretty conservative with my document features, too. I'm sure that many of mine would increase 10-fold at least. I do User-Agent-based serving and the variations are pretty subtle sometimes. For example, some browsers compress multiple tags down to one space and some don't. The specs *seem* to say that compression is correct behavior, but that's somewhat moot: there is currently a lot of variation in how available browsers interpret this. To get the layout I desire (let's not argue about conent vs presentation stuff, ok? it's been beat to death), I have page fragments that are correct for browsers that do compress and page fragments for those that don't compress. Using the proposed <INSERT> construct, I'd have to translate: Some text (OR Some text) to (roughly): <INSERT SRC="http://www.name.dom/uncompressBasedPage.html" TYPE="text/conditional-html"> <PARAM NAME="condition" VALUE="vendor-compressnbsp/1.0"> Some text </INSERT> This is a pretty big increase in the actual size of the file. Another disadvantage to this approach (as I started to realize in an earlier message I sent and as someone just pointed out) is that it's biased towards new browsers: old versions will never support the new construct. This is good insofar as it encourages browser vendors to conform, but it's trouble too 'cuz I can't do content negotiation with older browsers (as I can with User_Agent). >Also, how am I going to test each of the >different structures -- keep copies of the top five browsers for each >platform handy? Yuck! (Maybe my worry is just an artifact of the current >(lack of) automated / advanced tools?) BrowserCaps [see my .sig] helps, but doesn't truely allow for testing. ----------------------------------------------------- David Ornstein Outbreak: http://objarts.com/outbreak-unreg BrowserCaps: http://objarts.com/bc Personal Info: http://objarts.com/davido
Received on Monday, 29 January 1996 15:18:29 UTC