- From: Larry Masinter <masinter@parc.xerox.com>
- Date: Sat, 27 Jan 1996 16:08:43 PST
- To: wpk@fc.net
- Cc: www-talk@w3.org
William King wrote: > Is it not true that as long as folks ignore the central issue of > standardization of content and the provision of conformance testing tools > and branding programs for clients and servers, then they're going to get > bogged down trying to solve peripheral and unnecessarily complex problems > such as this one? Clearly, there are cases where content-negotiation is > obviously desirable, such as in the area of national-language-support. > However, focusing more resources on the conformance and branding problem > might just cause of lot of this other stuff to go away. The problems of content type negotiation with versions, features, and the like is something that has been an issue for as long as there have been computers, and doesn't seem to go away no matter what the forum. It isn't really even a branding problem. If you have a "Microsoft Word" file, you actually have to identify, in some situations, whether you have Word for Windows, Word for WIndows for Workgroups, Word for Macintosh, version 5,6, 7, etc. These are all the same brand, and they have some amount of interoperability. If you have a "TIFF" file, you need to know, in some circumstances, whether it is TIFF class B, class F, or uses non-standard tags. If you have a "Postscript" file, you might need to know whether it is black and white, color, what font set it assumes is on your printer, whether it's been paginated back-to-front, what paper size it assumes (A4 or US). These kinds of problems seem to be with us whether it is a completely closed specification (Microsoft Word), a published vendor specification (Postscript, TIFF), or that the specification has been handed over to a standards group (HTML) but there are proprietary extensions. Over time, some of the complexity in document formats goes away. Some of the enhancements get dropped, everyone finds a least common denominator to support, differentiation beyond that becomes unimportant, and stuff converges. There's a lot fewer image formats floating around now than there were 3 years ago, for example. (I'm surprised at the rise of PNG though.) In that sense, I agree with you that focusing more resources on the conformance and branding problem is productive, useful, and very important. On the other hand, the problem seems to be ongoing. As soon as we resolve all of the variations of feature sets for HTML, we'll discover that we have to deal with it for VRML. ("Now, does your VRML renderer do shading? Have fog-lamps on your headlights? etc.") So, I don't think that the issue of feature sets, characteristics, etc. for media types is going to go away, as long as there are new encapsulations of intellectual content ("document formats") to be invented.
Received on Saturday, 27 January 1996 19:09:02 UTC