Re: good practice for clients

On Tue, 18 Jun 1996, N.G.Smith wrote:
> >| I find it unbelievable that it isn't universally supported.
> >
> >Yep - in-lined infomercials must be a higher priority!
> This doesn't explain it. Browsers have been around for a lot longer
> than that. Why weren't the first generation of browsers built with this
> kind of support from day one.

The first generation (and second and third, etc, etc) didn't have this
sort of support in it because nobody was really doing the sort of serious
proxying that we're all up to now and there were much more pressing
problems taxing the minds of the then underfunded and underresourced
browser authors.  Spending lots of time on top notch proxy support would
have been pointless in the early days when people still struggling to
workout so much else in the browser.  These days we're all so used to
inlined images, background colours, forms, scalable fonts, blinking text
(:-)), etc, etc that we tend to forget that it was amazing to see
hyperlinks work at all when the early browsers were being hacked up out of
thin air by dedicated individuals and groups. 

Now we have a large market for browsers and lots of players committing
resources to it.  We also have a growing desire to have good proxy support
in them to make use of the new infrastructures we're all building (the
HENSA cache, the UK Harvest/Squid network, the US cache, NZ cache, etc). 
The commercial browsers like Netscape can keep up with this.  However its
not unnatural that the original under-resourced browser authors (where they
still exist) are having a tough time playing catch up.  That's probably
why Netscape has a <insert your favourite number between 50 and 90 here>%
of the browser market.

Of course older browsers with freely available source code have the
advantage that interested individuals can put in support for "wacky"
things like proxies (that don't look "cool" or even get noticed by most
users) and help out the authors a bit (cf: X Mosaic).  They're also great 
test beds to try out ideas and so will continue to be used be hard core 
weenies like myself.

> Something I've not had a chance to follow is the W3C libraries. If the
> LineMode browser has support for multiple IP addresses on a hostname,
> is the code not already there? What does Arena do for example?

I think the LineMode browser is built against a new common library than 
the CERN server (if what Martin was telling me the other day sunk in 
correctly).  Certainly when I was hacking proxying bits and bobs in 
Mosaic quite a while back, the common library in there was a heavily hacked 
over version of an older CERN library (it might have changed a lot since 
I last looked at it as Mosaic now contains all the proxying features I 
was hacking about with at the time).  As far as I remember, Arena was 
built against the latest release of the library so should Do The Right 
Thing(tm) if that functionality was part of the library.

Tatty bye,


Jon "Jim'll" Knight, Researcher, Sysop and General Dogsbody, Dept. Computer
Studies, Loughborough University of Technology, Leics., ENGLAND.  LE11 3TU.
* I've found I now dream in Perl.  More worryingly, I enjoy those dreams. *

Follow-Ups: References: