Re: Structured text v. page descriptions (was Netscape, HTML, and

Chris Lilley, Computer Graphics Unit (
Fri, 28 Oct 1994 04:06:46 +0100

In message <aad587ca0002100438a4@[]> Nick Arnett said:

> At 12:46 AM 10/27/94, Chris Lilley, Computer Graphics Unit wrote:

> >Main problem is, this is specific to a particular browser. I may well wish to
> >use others. I do not want to have to use another browser just to look at
> > links from pdf documents. I particularly do not want to have to use a
> > different platform.

> Could you clarify the first objection a bit? The second one I find quite
> reasonable. Having to buy a new machine to view a document would be a bit
> extreme. But I have a hard time imagining that it would be a good thing to
> try and make one browser, at least under the current architecture, try to
> do everything.

I said nothing about retaining the current architecture.

> Are you objecting only to the resulting inconsistent user
> interface?


> The memory footprint?

Well, I suppose

> The loose integration? (They're allrelated, I imagine.)

Yes, and yes they are.

The original posting talked about a special communication between a particular
PDF viewer and a particular Web browser - S-Mosaic. It has since transpired that
the Spyglass mods will be merged back into the public Mosaic code base, but that
does not alter the basic principle.

Suppose I am using some browser A. In the HTML document I am reading (1), there
is a link to a PDF. I select the link and up pops my PDF viewer. In this PDF
document is a link to HTML document 2. I select that, and up pops S-Mosaic to
show document 2. So now I have two web browsers up, one for my normal use and
one to show pages that happen to be referred to from PDF documents. This is
clearly (I hope) undesirable, no?

I can either go with this two browsers (two separate hotlists, two separate
history trails) approach, or more likely I would have to ditch browser A and use
S-Mosaic as my main Web browser. I do not want to be forced into this situation.

My comment about platforms was induced by finding that there is no Motif PDF
viewer for Unix/X platforms. Multi-platform does not just mean doing Macs as
well as WinDoze PC's.

> In Chicago, I made a joke about in-line PDF documents, which drew a hard
> stare and "Why not?" from Dave Raggett. I'm slow, but I figured out that
> his basic point, which is that just about anything might be in-lined, is at
> least worth considering.

Inline EPS - a closely related concept, I am sure you will agree - has been
feasible for years. In the days of Mosaic 1.1 I sent mail to Marc Andreessen
suggesting that X Mosaic might well use the Display PostScript X extension
supported by all major workstation manufacturers (bar HP ;-( ). I used to
use a mail reader that detected PS files and displayed them in this manner.

> One reason that the interface presented by Mosaic is consistent is that
> it's so limited.

I hope you are not going to confuse the conceptual issue of a single user
interface for all the hypertext internet communication tools with the specific
features that Mosaic for X has right now?

As I recall from old slides that used to be at CERN, TimBL concieved of the Web
on seeing scientists use 7 or so computer systems from several terminals, all
with different commands and interfaces.

I recall having to use xgopher, xarchie, mxrn, etc etc to access the various
internet services.

At the moment, a single browser can access a variety of servers (gopher, web,
news, ftp, etc) all with a single consistent user interface and these documents
can reference each other fairly freely. I can read an HTML document, select a
link to a news group, browse the articles there and select a link from one of
them to an FTP site or another HTML file, and mail the author. This is good.

I do not want the web to fragment such that I need my favourite browser plus the
one that can talk to Acrobat plus the one that does inline CGM plus the one that
handles sniff-o-matic URLs.

> In any event, today's situation is that we are coming from the helper
> application model, which could lead us toward a future in which Web
> browsing becomes a components of many applications, so the the browser you
> get depends on the nature of the document you retrieve. Some suggest that
> Mosaic, for example, might become little more than a window manager

Yes, that is what I was alluding to. Not only do I not want to need two or more
Web clients, I would prefer to have the document-like formats like PDF+Web,
HyperTeX, etc all handled by my browser. I speak from a user interface point of
view here. Whether the browser actually deals with a format or hands it off for
rendering to a helper application - much the best idea, IF the rendered result
is displayed by the browser - is irrelevant to me the user.

Your analogy of the browser being a windowe manager, or a shepherd for a flock
of cooperating applets, is a good one.