Re: LYNX-DEV DOS, DOS and Windows

[Rob Pelkey]

On Sat, 26 Oct 1996, Egor Egorov wrote:
> On Thu, 24 Oct 1996, Wayne Buttles wrote:
> > From my limited understanding of compilers, a true plain DOS port of 
> > the current Lynx2-6 is not possible because some of the obj files
> > generated are bigger than 64k.  If Lynx were restructured there might
> > be a chance, but a drastic change would take some co-ordination among
> > the masses.  It also generates a very big image--maybe too big to run
> > on a low end machine without optimizations.

Couldn't you get around this by splitting up the individual .OBJ files
into multiple segments (if you used the proper memory model)?  From my
understanding individual .OBJ files aren't what's limited to 64K, it's the
individual code/data/stack segments within those files that are limited -
I don't know whether you can break up code into multiple segments within
an .OBJ file, though.  My understanding of DOS compilers is rather
limited, not having done much above "Hello World" in Turbo C for
Windows...  (in console mode... :-)) 


> Look: Lynx is text-mode. Where now people can use text-mode www browser ? 
> 1st, for ergonomic websurfing.
> 2nd, if not enough memory.
> 3rd, for speed.
> ...
> ..., on low-end machines.
> I think, an old XT with slipper.exe or etherppp.exe with lynx will be very 
> nice to use!! So, I think, a realmode port is very important. maybe with
> some sex, but it have to be. :) 
> How d'you think ? 

Probably the only reason why I have an interest in the development of Lynx
is for the possibility of an 8-bit MeSsy-DOS version.  Our site has many
cubic yards of old 640K XT clones which would be perfect as kiosk
machines, were there a fairly up-to-date version of Lynx in existence that
would run on these systems; an 8-bit DOS version would fill some immediate
needs that I and others here have, in addition to providing browsing
capability for low-end users elsewhere (e.g. schools with poor funding,
non-profits on a shoestring budget, developing countries, etc.).

Of course, Lynx in its present form is simply too huge to fit into a 640K
box.  I think the thread a few days back about Lynx becoming bloated was
right on target - for friendliness to large multi-user systems where many
people run Lynx, and to make a DOS port a bearable task, Lynx has to be
more judicious about consuming memory and other system resources.

If an 8-bit DOS port of Lynx is ever to reappear, then a 32-bit or 16-bit
DOS port is a step in the right direction; once that's done, then a few
brave and committed souls can perform the brutal hackery necessary to fit
Lynx into an 8-bit, 640K memory model.  (Ugh...)  Odds are that I may be
one of them. 

Comments?

                                                        Rob

--
<!-- Rob Pelkey <address@hidden> \ http://www.bates.edu/~rpelkey/ -->
<!--  Professional Web Dude          \  #include <std-disclaimer.h>  -->
<!--   Math Major, Bates College '98  \  Join the march to save      -->
<!--    News Director, WRBC 91.5 FM    \  individuality!             -->


;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;