|
|
| Author |
Message |
| 25 new of 109 responses total. |
bellstar
|
|
response 25 of 109:
|
Jul 22 09:11 UTC 2008 |
I generally am not very relevant in this discussion but I'm posting this
anyway.
CMSs are "certainly" not the way for Grex to go. Their functions only slightly
overlap with those of Grex's and despite what remmers says they are far from
being "easy to maintain." They are almost invitations for disaster because
they're too complex, often involve many "moving parts," and mostly hinge on
"shiny new technology" instead of "proven old ways."
A simple CMS package will need heavy server-side (e.g. PHP, Perl) and
client-side scripting, some sort of database (e.g., MySQL, PostgreSQL), and
"mods" for the web server (e.g. mod_php for Apache). These are obvious
performance drains and security threats. Grex is known (and loved for) being
simple and "old-skool." Some Grex users delibrately use very minimal setups and
view these same pages through lynx (or through bbs), with those "frills"
they'll be deterred. I like it when pagefuls of good talk is loaded over my
dial-up connection in under 15 seconds.
On the functional aspect, Grex offers a bundle: shell access, email,
webhosting, "and" community interaction. CMSs are focused only on collaborative
content creation which is not the sort of community interaction Grexers are
involved in. Grex isn't hosting some sort of "project" that needs
"collaboration," i.e. modification of the "same" content. Its community
interaction paradigm is very well described by the title it already has: bbs.
The refurbishment it needs in that respect is probably some new "light"
bulletin board software but I'd be skeptic to even deploy phpBB and the like
because all of them, like CMSs, depend on server-side scripting and DBMSs.
As for "standards compliance," as far as I know most "recent versions" of
standards have been specifically designed to be supersets of their previous
versions. So Grex is not non-standard if it still presents content in old HTML
instead of some blasted XHTML/DTHML 4.1-bloody-transitional. In fact, one has
to be proud to present a functional Grex using the smallest possible subset of
the pertinent standards. That's good minimalism.
I believe on a "frills-mostly-with-a-bit-of-content" Internet a good ol'
system's mission should be to keep up the good ol' ways. In case of Grex, that
could be accomplished by enriching the shell/remote access experience (for me
that boils down to port forwarding and running httptunnel ;-), expanding (or
total abandoning) of email services (to include IMAP, for example), relaxing
webhosting practice so that people can make diverse websites within reason (I
don't think people should be able to put up sites written in PHP on Grex,
though), and a more capable but still "text-oriented" bbs (a "multi-level
forum" structure instead of conferences, and a proper linkable archive).
Good luck to [put the hero's name here] who will revive Grex. Seriously though,
I'll be happy to help if there's something I can do for Grex over some
thousands of kilometers and a dial-up connection.
|
keesan
|
|
response 26 of 109:
|
Jul 22 13:35 UTC 2008 |
Perhaps fronttalk could be changed so that you can click on a link posted in
bbs, and use lynx or links to view it (no graphics).
|
remmers
|
|
response 27 of 109:
|
Jul 23 15:01 UTC 2008 |
Re resp:25 - Hmm. I'll need to educate myself better on CMS's before I
can agree or disagree with what you say. As far as the underlying
plumbing is concerned - Perl and MySQL we already have, and intelligent
use of CSS and PHP could substantially improve the website. I've found
that adding some light-weight CSS and PHP to my personal website has
significantly simplified maintenance tasks. I'm in favor of reasonable
minimalism and avoiding technological overkill, but not to the point of
rejecting technologies that could be useful and make our jobs easier.
Re resp:26 - Adding mouse support to a tty interface is an interesting
concept, but probably not feasible (although xterm has limited support
for it, I think). On the other hand, implementing a web interface that
looks and acts substantially like the tty interface, but also supports
"clicking", might have some appeal to the old guard. Not that I'm
seriously suggesting that it be done.
|
keesan
|
|
response 28 of 109:
|
Jul 23 15:29 UTC 2008 |
You don't need a mouse to click on a link. Pine lets you go to URLs already
(Enter key). Why not fronttalk also letting you invoke lynx or links or even
w3m?
|
remmers
|
|
response 29 of 109:
|
Jul 23 15:35 UTC 2008 |
Ah, gotcha - "selecting" might be a less confusing term than "clicking",
which suggested to me that you meant a real point-and-click device
(mouse). Well, it's definitely feasible technically; stumbling block
would be, as usual, persuading someone with the necessary programming
skills to implement it.
|
keesan
|
|
response 30 of 109:
|
Jul 23 16:02 UTC 2008 |
My Enter key clicks.
|
cross
|
|
response 31 of 109:
|
Jul 23 16:20 UTC 2008 |
resp:28 I'm not sure what the point would be.... Grex already has
fronttalk that works with a TTY interface. One of the ideas is to
make the web interface more TTY like, but that's a separate issue.
|
remmers
|
|
response 32 of 109:
|
Jul 23 17:32 UTC 2008 |
The point would be to make it easier in Fronttalk to follow web links
that people post, similar to a mechanism that already exists in Pine.
That would actually be a nice facility to add, if somebody gets inspired
to do it, and underscores why Grex really needs to be running open
source conferencing software.
|
tod
|
|
response 33 of 109:
|
Jul 23 17:35 UTC 2008 |
I wouldn't mind a web interface to BBS which supports cellphone users.
|
cross
|
|
response 34 of 109:
|
Jul 23 17:50 UTC 2008 |
resp:32 Aha. Okay. It strikes me that both picospan and fronttalk
pipe conference data (read: responses) through an external pager
program; I wonder if less or most of any of those can be made to
recognize URLs and do something similar to what pine does. If that
were the case, we could just set the pager and get the same
functionality for free.
|
katie
|
|
response 35 of 109:
|
Jul 24 01:54 UTC 2008 |
The new page looks nice. A couple of typos: "frequently" needs an "l."
"Our home, Ann Arbor" needs a space.
|
cross
|
|
response 36 of 109:
|
Jul 24 02:34 UTC 2008 |
Thank you! I corrected both typos; please let us know if you see any more!
|
bellstar
|
|
response 37 of 109:
|
Jul 24 04:37 UTC 2008 |
Re #27:
I'm rather bewildered to hear Grex runs a SQL server. Are there any Grex
applications depdendent on it? Has it ever been load tested? Or is it just
"educational?"
As for Perl, mere availability of the interpreter is not enough. PerlCGI
modules must be added to both Perl and Apache (assuming Grex is served on
Apache) to provide the required CGI hooks. These hooks and the "possibility" of
remotely running malicious code through a misconfigured or buggy CGI interface
will make Grex vulnerable to a host of attacks. On the other hand, cgi-bin is
already there so perhaps there won't be much added vulnerability.
I've heard computer security experts have a saying that goes, "usability is
inversely proportional to security." I wonder if Grex isn't already "useful"
enough?.
|
nikm
|
|
response 38 of 109:
|
Jul 24 09:23 UTC 2008 |
The front page is too cool. But the same theme is not followed in all
pages. It should be consistent in all pages as remmers mentioned. I
would like to help you in always I can. I can read/write HTML, CSS,
little PHP and Perl.
Thanks
Noorul
|
cross
|
|
response 39 of 109:
|
Jul 24 11:16 UTC 2008 |
resp:37 I've found it's not always true that security and usability
are inversely proportional. For instance, many networks become *more*
usable when moving to, e.g., Kerberos for authentication because it
suddenly becomes transparent to most users: instead of entering a
password/passphrase for a variety of different services, the library
just pulls it out of your credentials cache on your behalf. Similarly
with SSH public-key authentication.
Btw: Grex runs MySQL for a few different, administrative types of
things. We also already run mod_perl.
resp:38 Thanks! That'd be great! All of the HTML is
in /var/www/htdocs; feel free to copy that to your directory and go to
town making changes that you see fit. I agree that the pages should
have a consistent look and feel, but haven't had time to update most
of them myself.
I *did* convert the HTML in the main directory to XHTML using tidy,
but it's being servied up by Apache as text/html instead of the
XML+XHTML combo that all the kids love. That's largely a matter of
fixing links.
|
remmers
|
|
response 40 of 109:
|
Jul 24 14:14 UTC 2008 |
I seem to remember that IE6 has problems with the application/xml+xhtml
MIME type.
|
cross
|
|
response 41 of 109:
|
Jul 24 14:51 UTC 2008 |
It does.
Which is why I modified the Apache configuration to rewrite the MIME
type to text/html for .xhtml documents if it detects that it is
serving to Internet Explorer. Similarly for Lynx.
For other browsers, we still want to serve as application/xml+xhtml.
|
hera
|
|
response 42 of 109:
|
Jul 25 13:35 UTC 2008 |
I still hate the colors.
|
mickeyd
|
|
response 43 of 109:
|
Jul 25 22:50 UTC 2008 |
And what happens when STeve finally logs in realizes things have
changed? Will he / can he revert back to the old and lock cross out?
Only 1/2 joking..
|
lar
|
|
response 44 of 109:
|
Jul 26 02:15 UTC 2008 |
I say we give cross root and lock steve out. It would probally be
months before he noticed anyway.
|
hera
|
|
response 45 of 109:
|
Jul 26 15:17 UTC 2008 |
Change the fucking colors!! It looks like diarrhea.....like someone
hershey-squirted all over the screen.
|
cross
|
|
response 46 of 109:
|
Jul 27 01:36 UTC 2008 |
resp:32 Once again, we mention the need for open source conferencing tools.
I think we need to push strongly on this; maybe we should get with the M-Net
people and collaborate on something new.
|
remmers
|
|
response 47 of 109:
|
Jul 27 18:21 UTC 2008 |
Shortest path to something useable would probably be Backtalk +
Fronttalk. Backtalk works well, and fixing up Fronttalk would be a lot
less work than developing something new.
|
cross
|
|
response 48 of 109:
|
Jul 27 19:32 UTC 2008 |
Well, I think the absolute shortest path would be getting either Marcus
or the Thalers to open up the source for either Picospan or YAPP,
respectively. However, I'm not really sure what either partys' positions on
the matter are. At the minimum, with respect to Picospan, it sounds like it
would require a lawyer to figure out the ownership of the source code. If
Marcus were interested, I wonder if someone like Dave Cahill might be
willing to look into that on a pro-bono basis for Grex.
But, history is usually a strong indicator of the future, and it seems
extremely unlikely that either YAPP or Picospan will be open sourced any
time soon. Given that, we need to move to plan B.
And on that front, I absolutely agree that backtalk + fronttalk is the
path of shortest resistance to getting an open source replacement for
picospan on Grex. The problem now shifts to finding the resources to get
fronttalk cleaned up to the extent that it will make a credible replacement.
Fronttalk is ... not a small program. It's not *big*, but it's not
trivial by any measure. And it's not well understood except by Jan, who's
mostly not around these days. And we don't have enough people testing it.
What to do?
Well, it's open source, so if we can get some more peopple who are
decently familiar with Perl and how Backtalk works together, we can work
through the bugs easily enough. I mean, nothing here is particularly
complex; there are no deep mysteries of computer science. It's just
maintenance on a decently sized program. But where are those people going
to come from?
Another option is to try and persuade Jan to work on it again; I suspect
we'd have to pay him (as I suspect part of the reason he hasn't done it
already is that he's busy doing things that make enough money to help him
support his family). Purely to get some information, I sent him an email to
ask him if he'd be willing to do some work if we paid him. However, he
hasn't gotten back to me (or, if he did, it got flagged as spam by GMail and
I never saw it). I suspect, however, that we haven't really got enough
money to make it worth his time.
So we are back to doing the work ourselves. So who is willing to take
up the challenge of working on fronttalk to turn it into a suitable
replacement for Picospan? Note, you don't have to be staff to do this, you
just have to know some Perl.
And who is willing to help out testing it? We *really* need people
to *use* it and tell us where the problems are. And if we decided to make
this change, it's likely that people might have to change some of their
.cfonce files and maybe some minor usage patterns. Are people going to
accept this? Will they work with us to get everyone converted over to the
new software, even if it's slightly incompatible with Picospan?
|
denise
|
|
response 49 of 109:
|
Jul 28 01:50 UTC 2008 |
I'd be willing to help but I'm not literate in computer technology...
|