- From: Peter Flynn <pflynn@imbolc.ucc.ie>
- Date: 28 Aug 1997 22:10:37 +0100
- To: Albertfine@aol.com
- Cc: www-html@w3.org
Albert Fine writes: I think I see a flame war simmering so let me make a few things clear. I don't work for Double Click :) I am a computer consultant. I have taken classes on HTML and I have reviewed the XML and HTML 4.0 specifications. I am sure everyone here knows something about HTML. My ideas are very early in development. HTML streaming basically; organizes the tags in a HTML document to send what is displayed first. Tags that are not displayed such as keywords, signatures etc. would be added at the end of the file. OK, there's no reason why you shouldn't rewrite the HTML DTD to do this. But it wouldn't then be HTML...but it might be called SHTML (Streaming HTML). BTW they're not called tags: they're called elements. Tags are the things that mark the start and end of elements. file. These protocols would probably be used by HTML editors. HTML streaming also involves two tags; the events tag and the stream tag. The event tag would list all the tags in the HTML file with rough descriptions so the browser knows what to expect. The stream But the DTD already does this, and software already exists to handle it: why would you want a separate file/stream? tag would tell the browser you can display this now, it will not be changed. For example, someone mentioned the problem with loading tables. The HTML editor could calculate the impact of size elements in the table tag. It would add a stream tag to points in the table. The stream tag would tell the browser to display the unfinished table knowing that it will not be changed. Or the HTML editor could This already exists in some HTML DTDs: it's called COLSPEC or COL. It was designed specifically for what you describe. add a rough description of the table to the events tag. You would have the table streamed without major changes to the table tag or using a proprietary tag. I don't think tables are the last or only example of the need for HTML streaming. I don't think the speed at which a page loads while downloading is the biggest problem in the world but I do think we should take the initiative. Please review all the articles in the thread; http://lists.w3.org/Archives/Public/www-html/1997Aug/thread.html#290 I'm still not clear what needs to be done to enable streaming that cannot already be done with existing definitions. I don't think that removing the header to the end of the file is productive merely on the grounds that the information in it is not displayed: there are other reasons for things to occur where they do -- indexing engines for example want to find the list of keywords near the top of the file, to minimize the volume of data they read before getting to them. I think you'll find that the organization of HTML -- however deficient in other areas -- was done for good reasons, and I'd hesitate to undo it unless for equally good ones. It does strike me that you have a very good potential application, but that it could be implemented much more easily by using all the existing facilities first, before you add new ones or change things around. ///Peter
Received on Thursday, 28 August 1997 17:14:03 UTC