Subject: Re: Suggested change to README.html generation: tables for binaries
To: Nathan J. Williams <nathanw@MIT.EDU>
From: None <mcmahill@mtl.mit.edu>
List: tech-pkg
Date: 11/07/1999 09:21:11
On 6 Nov 1999, Nathan J. Williams wrote:

> 
> I was looking at the README.html for a bunch of packages, and found
> the list of binary packages to be somewhat annoying to use. It's long,
> and not sorted by either version or architecture. I figured that a
> table with versions vs. architectures would be a good way to do this. 

I had been thinking the same thing.


> The big downside I see to this patch is that it makes the pages much
> less readable with non-table-aware browsers (Lynx, for example). I

yeah, this is a fairly big downside.  

> don't know what we expect the audience of this HTML to be,
> really. It's been suggested that the people who look for binary
> packages on a web page and the people who use non-table-aware browsers
> may be completely disjoint sets.

I'd disagree.  Some of my machines are pretty slow which makes lynx nice
(plus, on pmax, mac68k, etc, there isn't a decent netscape), its also
those machines when a quick binary install is quite nice as opposed to a
long build.  For example, scilab takes 17 hours to compile on my Mac IIci.
It took me several days to build the packages I usually use on my sparc
IPX.  


So, I agree that under netscape your README looks really nice and I like
it a lot, I'd just like to see something which works ok with lynx.  

visually, its almost there under lynx anyway, the packages are sorted by
arch and then by version, you just can't tell which versions you have...

what about organizing it like:


alpha:   
	1.4     mpg123-0.59q

amiga:
	1.4     mpg123-0.59q
	1.3.3   mpg123-0.59o
	1.3     mpg123-0.59k

etc..

its still somewhat long vertically compared to your table, but is sorted
in a much easier fashion than the current approach.

-Dan