- From: nemo/Joel N. Weber II <devnull@gnu.ai.mit.edu>
- Date: Sat, 22 Mar 1997 17:54:22 -0500 (EST)
- To: BruceLeban@aol.com
- CC: www-html@w3.org
From: BruceLeban@aol.com Date: Sat, 22 Mar 1997 17:00:05 -0500 (EST) For another perspective on "HTML should be an output format", check out Globetrotter Web Publisher designed with that philosophy in mind: http://www.akimbo.com/globetrotter Of course, Globetrotter writes HTML right now because that's what the web uses, but if the web were to change overnight to a new language, Globetrotter users would just republish (as easy as reprinting a document) and go. (After of course upgrading to the new version of Globetrotter that writes the new output format.) A couple points: 1) There is no way the web would change overnight. From what I can see, HTML will probably always be needed for backwards compatibility. 2) Globetrotter has to store the document in some format, right? My opinion is that HTML is essentially a standard, open format. What happens if I write something in Globetrotter today, and ten years from now I have some Globetrotter files, but I no longer have a computer that can run Globetrotter? I actually have the problem of converting AppleWorks files to something that my Intel machines can read. Moving the files to a DOS disk is not a problem. Since I have documentation of the internal format AppleWorks uses, reading the files shouldn't be a problem. But then there's the question of what format to generate. Plain text will work fine, but it destroys the formatting. I could use TeX or HTML as the output format easily enough. But what would really be optimal for the current use is creating files in the native format of a particular proprietary word processor, for which I have no documentation. And I wonder what I'll have to do another ten years from now. (I've become a free UN*X user; but for some reason the rest of my family insts that UNIX is too hard to use.) So I conclude that something standard like HTML is a great storage format. In particular, Globetrotter rejects document=web page. After all, no one edits word processing documents with each page in a separate file. Can you imagine anyone believing that was the *right* way to do it? A single Globetrotter document (1 file) can publish many different HTML pages (many files) on the web. That's not really a valid analogy. One page = one file on the web because a web page can be as long or as short as is desired. There's a compelling reason to not have one file corrrespond to one printed page: if I insert another word in the middle of my document, I want words to spill over to the next page automagically. Keeping the whole document in one file makes sense for this. But I don't see why one file == one page ever causes lossage on the web. Apologies if anyone thinks this message is too self-serving. We've been trying to promote the "HTML should be an output format" message for quite some time and it's difficult. Most people just don't seem to get it. Most of the people publishing documents on the web are the early adopters who have been forced to learn HTML. There's some reluctance for them to accept that there might be easier ways to do things, that would deny them their web"master" status. My personal frustration is how many people are impressed with tools that basically just type < and > for you. I agree that many web page editing tools are not as useful as some people think they are. But I don't see why using HTML as an internal format causes problems. I also get frustrated by people who don't understand the underlying technology and are confused when any lossage happens. HTML is not that hard. Let people learn it. Then we don't have problems with slightly broken tools. (Admittedly the tools should be fixed. But how do you let people intellegenetly choose between relative and absolute links without explaining them?)
Received on Saturday, 22 March 1997 17:54:56 UTC