You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Brian Behlendorf <br...@organic.com> on 1996/01/16 05:04:24 UTC

The List

Hmm, I'm sensing the development fires to be cooling somewhat.  Perhaps 
if we had a list of projects to focus on, we'd be better off?  In that 
hope, I wrote up into an HTML file a list of things I've been scrawling on
various scraps of paper up to this point, to see if perhaps we could 
prioritize the work we do so we don't get overwhelmed with the number of 
things on our plate?  

A stab at that is at http://bong.com/httpd/todo.html.  It's split up into 
a list of known bugs (at least the ones I could remember), things we have 
agreed would be a Good Thing to put on the to-do list, and finally "wish 
list" items.  I'd like to first ask for submissions for additional 
entries on the list, and then after that we can talk about priorities.  
Howzatsound?

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  brian@hyperreal.com  http://www.[hyperreal,organic].com/


Re: The List

Posted by Alexei Kosut <ak...@nueva.pvt.k12.ca.us>.
On Wed, 17 Jan 1996, Brian Behlendorf wrote:

> Since you've poked at it more than the rest of us, could you make a stab 
> at least at what such an api might look like?  I don't see why an MD5 
> module needs to wait for that, unless you're concerned about N different 
> authentication methods proliferating into N different modules...

It doesn't, I guess. But that is what I'm concerned about. At any rate, I
really am not familiar enough with the facets of C needed to generate an
auth API, someone else can do it, but I guess I'd be more than happy to
write a mod_md5 stand-alone. Probably with just flat-file and DBM support.
I'll do that. 

> Actually, the content negotiation subgroup isn't going to have a huge 
> list of changes - as defined in HTTP/1.1 content negotiation is fine for 

Okay... I wasn't sure about that, and I figured it was probably safer to 
wait until any implementation beyond what we have now is done.

--/ Alexei Kosut <ak...@nueva.pvt.k12.ca.us> /--------/ Lefler on IRC
----------------------------/ <http://www.nueva.pvt.k12.ca.us/~akosut/>
The viewpoints expressed above are entirely false, and in no way
represent Alexei Kosut nor any other person or entity. /--------------



Re: The List

Posted by Brian Behlendorf <br...@organic.com>.
I've added everyone's comments to the list, and added a set of initials 
after each to denote who made the comment, to refer to later in case we 
forget why something was suggested. :)

On Mon, 15 Jan 1996, Alexei Kosut wrote:
> 1. Regarding /~, my mod_userdir patch can do that, and is in for_Apache_1.1b0

Ah, right, okeydokey.  I support making this a part of the core 
distribution.

> 2. I poked a little at MD5, but I'm not sure I can get a chance to follow 
>    through on that, especially if we want to go to an authentication API 
>    (which I think would be a good idea).

Since you've poked at it more than the rest of us, could you make a stab 
at least at what such an api might look like?  I don't see why an MD5 
module needs to wait for that, unless you're concerned about N different 
authentication methods proliferating into N different modules...

> 3. I'm not sure a graphical installation process is needed. What does it 
>    take to install? Not much. Maybe edit a line in Configuration, Configure, 
>    and make. We throw in some sort of automatic configurer like Gnu 
>    autoconf, and it's one step. I think a graphical process would be more 
>    trouble than it's worth. 

Yeah, the graphical bit was really more intended for the second item, the 
administration service, but still something that made initial 
first-ever-user setup and configuration easy would be a Good Thing.  
Right now it's a mishmosh mostly because of history and feature creep, I 
think it could be cleaned up a leetle.  

> On the other hand, a graphical *adminstration* 
>    process would be nice. However, that *should* be done via the web, 
>    because it does make remote administration very easy. Not all of us have 
>    access to X.

Here I'm inclined to disagree - while nice in theory, in practice (and 
I've administered two different servers with this functionality) it 
feels very "remote" to me.  Sorta like sitting on the beach watching 
surfers for 3 hours then telling your friends you went surfing.  
Something like that.  Sure, with server-push and client-pull that 
"dynamicism" can be emulated, but... I'm not making a great case here, 
you'll just hafta trust me :)  Keep in mind there are free X servers for 
the Mac and Windows, and the X consortium has Broadway up their sleeve 
(an X server implemented as a Netscape plug-in).

Side note - why doesn't Netscape just become an operating system and be 
done with it?  Sorry...

> 4. I'd hold off on the HTTP/1.1 stuff (content negotiation especially) 
>    until the http-wg comes back with something more solid. While certain 
>    things (like Keep-Alive) aren't likely to change, some things are 
>    almost certainly going to.

Actually, the content negotiation subgroup isn't going to have a huge 
list of changes - as defined in HTTP/1.1 content negotiation is fine for 
the granularity of content-types and levels.  The HTML feature set 
problem is not going to be solved by us, in fact manmy of us feel it's 
something that should be solved in HTML using conditionals if possible. :)

> Oh, and toss a proxy module onto the pile somewhere. I'm almost done with
> one *grin*. By next weekend, I should have a working version you can all
> play with, though I make no gaurantees. 

Grrrrrrreat!

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  brian@hyperreal.com  http://www.[hyperreal,organic].com/


Re: The List

Posted by Alexei Kosut <ak...@nueva.pvt.k12.ca.us>.
On Mon, 15 Jan 1996, Brian Behlendorf wrote:

> Hmm, I'm sensing the development fires to be cooling somewhat.  Perhaps 
> if we had a list of projects to focus on, we'd be better off?  In that 
> hope, I wrote up into an HTML file a list of things I've been scrawling on
> various scraps of paper up to this point, to see if perhaps we could 
> prioritize the work we do so we don't get overwhelmed with the number of 
> things on our plate?  

Maybe. Although, I'm still wondering whatever happened to 1.0.1. Can
someone please tell me?

> A stab at that is at http://bong.com/httpd/todo.html.  It's split up into 
> a list of known bugs (at least the ones I could remember), things we have 
> agreed would be a Good Thing to put on the to-do list, and finally "wish 
> list" items.  I'd like to first ask for submissions for additional 
> entries on the list, and then after that we can talk about priorities.  
> Howzatsound?

Hmm. Looks like a good list. Some notes on the items:

1. Regarding /~, my mod_userdir patch can do that, and is in for_Apache_1.1b0
2. I poked a little at MD5, but I'm not sure I can get a chance to follow 
   through on that, especially if we want to go to an authentication API 
   (which I think would be a good idea).
3. I'm not sure a graphical installation process is needed. What does it 
   take to install? Not much. Maybe edit a line in Configuration, Configure, 
   and make. We throw in some sort of automatic configurer like Gnu 
   autoconf, and it's one step. I think a graphical process would be more 
   trouble than it's worth. On the other hand, a graphical *adminstration* 
   process would be nice. However, that *should* be done via the web, 
   because it does make remote administration very easy. Not all of us have 
   access to X.
4. I'd hold off on the HTTP/1.1 stuff (content negotiation especially) 
   until the http-wg comes back with something more solid. While certain 
   things (like Keep-Alive) aren't likely to change, some things are 
   almost certainly going to.

Oh, and toss a proxy module onto the pile somewhere. I'm almost done with
one *grin*. By next weekend, I should have a working version you can all
play with, though I make no gaurantees. 

--/ Alexei Kosut <ak...@nueva.pvt.k12.ca.us> /--------/ Lefler on IRC
----------------------------/ <http://www.nueva.pvt.k12.ca.us/~akosut/>
The viewpoints expressed above are entirely false, and in no way
represent Alexei Kosut nor any other person or entity. /--------------




Re: The List

Posted by Mark J Cox <M....@bradford.ac.uk>.
> I'd like to first ask for submissions for additional 
> entries on the list, and then after that we can talk about priorities.  

* Bugs *

<li>	Fix problem with CGI ignoring multiple same-name headers (combine them)

* ToDo *

<li>	Fix Cookie Module so it doesn't interfere with other site cookies
<li>	Allow "ErrorDocument" anywhere (One line change but needs user 
		configuration)
<li>	Clean "Shambhala" messages in source
<li>	Nicer Access Control (allow "OneOf","AND","OR" ala NCSA)

* Wish List *

<li>	Allow Byte Ranges for completion of interrupted transfers (Netscape 
		2.0b4+ client seems to have byte-range code in it)
<li>	Give a menu of options on Imagemap when used with Text browsers
<li>	Integration of a scripting language like PHP
<li>	Log unknown  HTTP headers received (so we can see when Netscape
		enable Byte-Ranges etc.)
<li>	Clean up source/comments still left from NCSA/1.3

* Docs For *

<li>	Cookie Module
<li>	DBM documentation: explain about having username/group in same file

Mark
Mark J Cox,  mark@awe.com -------------- URL:http://www.awe.com/mark.html
University of Bradford, England ---------- tel +44.1274.384070/fax 391333