W3C home > Mailing lists > Public > www-talk@w3.org > September to October 1996

Re: HTML questions ?

From: Murray Maloney <murray@sq.com>
Date: Mon, 16 Sep 1996 12:01:36 -0400
Message-Id: <>
To: jna <jna@retina.net>
Cc: Murray Maloney <murray@sq.com>, Eugene Venter <VENTERE@telkom.co.za>, www-talk@www0.cern.ch
At 02:50 AM 10-09-96 +0500, jna wrote:
>On Tue, 10 Sep 1996, Murray Maloney wrote:
>> At 03:55 AM 09-09-96 +0500, jna wrote:
>> >On Mon, 9 Sep 1996, Eugene Venter wrote:
>> >> I'm currently using Hotdog v2.53 as my HTML editor and the MAILTO
>> >> function is limited here.
>> >Ergh, you should go and read the HTML spec instead of trying to code with 
>> >a silly gui-based html builder. Use a text editor, _learn_ the spec, and 
>> >apply it. You will be limited if you allow all of your knowledge of html 
>> >to come second-hand from the people that wrote the editor. 
>(Okay, so maybe this was a teensy bit unfair, but...)
>> That's not quite fair. SoftQuad has participated in the 
>> development of the HTML spec since 1993. Our product,
>> HoTMetaL PRO includes the most up-to-date HTML implementation
>[... various product information deleted ...]
>I can completely understand the fact that you wish to promote your 
>product as being fully compliant, and that you are standing behind your 
>product in every way (including working with the appropriate groups to 
>make sure that you support all of the variants of html) , but there is 
>something to be said for learning the specs yourself. I'm not attacking 
>your product in particular, but I'm trying to make a stand here and say 
>that people should research the specs and learn what they're dealing with 
>before they begin to code. Most people nowadays pick up an 'HTML for 
>dummies book with free html editor inside' and just start churning out 
>broken HTML. 

Sure I am standing behind my product, but that's not all.
I agree that it is useful to learn a spec if you are going
to be working with it all the time. If you aren't, it is
probably a good iead to work with a tool that generates
"good" HTML.  
>Maybe it's that noone wants to learn anymore, and people want to just use 
>tools to generate things... It's a common argument in the GUI and non-GUI 
>crowds, non-GUI users insist that the GUI users don't want to learn, and 
>the GUI users insist that the 'computer is a tool, and just should work 
>like one'. Yet another religious argument that is completely out of the 
>scope of this list.

I doubt that "nobody wants to learn anymore", but I bet that
most people have way too many other things to do. For a large
percentage of computer users -- my wife and teenage daughter
included -- there is simply no need to learn a how something
works so long as it does work. I happen to agree that, for them,
this is the right approach.
>Let us give your product the benefit of the doubt and say that every 
>single tag known to man, made up by netscape, or developed 2.5 days ago 
>by yet another company fighting for absolute control of the standard 
>exists in the editor. Now, for the editor to be effective, it has to give 
>lots and lots of warnings and information about the particular grammar of
>the language. Most do not. Most just allow you to highlight something, 
>and say "Oh, let's make this a Heading #1" and it auto-inserts the tags 
>with no intelligence or attention to grammar, so it's technically 
>possible to 'pile up' commands so a line such as:
><H1> This is a heading with <I>italic</I> text </H1>
>is possible. Although, the spec says it's completely wrong to make a 
>header in this manner, people do it, and Netscape/MSIE supports the 
>displaying of the (originally) invalid html. 

Actually, the spec does not say that this is illegal, but back 
to your earlier point about warnings and grammar. My company's
product is an SGML-based editor, so it is guided by a very 
strict grammar for HTML. As a matter of fact, it is impossible
to create an incorrect HTML document in HoTMetaL. It is possible
to create an HTML document that uses extensions from Netscape,
Microsoft, W3C and others, but HoTMetaL provides you with a way
to "validate" the document and be informed that you are using extensions.

>How are users supposed to know how certain things are rendered in 
>differening environments if they only rely on the tags that are given to 
>them in the editors? They have to go out and test their documents under 
>each system, like most of us do, but the majority of people producing 
>HTML just outright assume that their audience is going to be 100% 
>netscape, or 100% MSIE and code for it, leaving thousands of people out 
>in the cold, and information inaccessible to users that are using textual 
>browsers, or older browsers like Mosaic or AOL

The only way to know how things will be rendered in different viewing
environments is to view them there. HTML does not prescribe format.
In fact, the people who have helped to write the HTML specs -- along
with Tim BL and Dan Connolly -- have been very careful not to specify
the formatting of HTML elements, but only to make suggestions.
>We write 95% of our HTML using emacs, and only switch to GUI based tools 
>when we need to convert large-scale documents, then passing the documents 
>from both of these sources into weblint-like verification tools that were 
>produced by the W3 consortium. We also try very hard to code for all 
>browsers, and not to be inconsiderate of the thousands of users out there 
>who still don't have the 16 megs of ram and disk space or T1 to use 
>netscape or graphically intensive pages. 

I wish that more web authors were as considerate and methodical.
Unfortunately not all are, and some cannot afford to be.


Murray Maloney				murray@sq.com
Technical Director              	http://www.softquad.com
SoftQuad Inc.				Phone: +1 416 544-9000 x2219
20 Eglinton Avenue West, 12th Floor	Fax:   +1 416 544-0300
P.O. Box 2025				
Toronto, Ontario, Canada  M4R 1K8
Received on Monday, 16 September 1996 12:05:16 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:32:59 UTC