Re: lynx-dev Lynx lite

[Nelson Henry Eric]

> If there is a clear consensus among the people who would want a
> lite image as to what capabilities could be pruned, then it makes

OK!  Al, you found the perfect name!  LITE_IMAGE is what it'll be.

> sense to use a single ifdef switch.  Otherwise it may make more
> sense to do a two-step process and first make sure that there is

I don't plan on doing anything quite so heavy duty, even if I could.
There probably is no chance for a "clear consensus", but it is my
belief that a reasonable compromise can be reached.  There are few
if any people besides me who are interested.  There was the fellow
who put Lynx on a hand-held device (bar-code reader?) and the minix
person, but they never bothered to come back AFAIK.

> Somewhere in the mists of time I recall discussion of libraries
> and dynamic linking as memory savers.  Is this of any help in

It is my understanding that this only decreases the size of the
physical binary which sits on the disk.  The image loaded into
memory at run time is the same??

> What is the performance equation you are fighting?  is it
> 
>   users = fixed_memory_size / (lynx_size + const.)
> 
> where the system gets unusable when there are more than 'users'

If it were only Lynx, there wouldn't be a problem.  It's toward the
end of the second semester when everyone has learned about all these
neat programs and it seems like they've got every program on the
system in memory.  I jumped from 12+36 MB (main/swap) to 84+160 MB
this Spring, but I was forced at the same time to run my own bind,
sendmail, httpd, etc. in order to gain the autonomy to do the upgrade.

So I apologize if this thread is too far off from lynx-dev.  I DO
appreciate very much the input and may in the future ask for advice
from the list.  Thank you all for bearing with me.

__Henry