Being a real maniac programmer I wrote CGIs for the site's news system
in C++. Sure, not the best way to accomplish such a task with, and I
would prefer perl. But unfortunatelly, the latter lacks such a great
instrument like the parser
library written by one of my ex-colleagues at NIX Solutions. It allows
to define and process templates of quite a nice kind. As to PHP, haven't
manage to put my hands on it yet. Anyway, now why I started with telling
you about all of that. Last week the news system suffered some changes
and improvements. For example, now it's possible for me to answer
visitors' comments for news items. Responses are shown next to comments.
Also a copy of my reply can be received with e-mail if you wish, there
is a check box in the comment add form.
Hopefully the improvements will increase the comment writing activity of
web site visitors. The situation is really a bit strange, for the
statistics says the news section is quite frequently visited. Sometimes
people ask me when I gonna write something new, or refer to my notes
here in e-mail or with icq. But there is nothing but hearing your
friends quoting texts from your own site. So I conclude this proves
there is an interest. However people seem to come to read something, and
not to communicate.
Another thing. Modifying the templates I made a little personal
discovery. The point is that I had never used CSS before, and it
appeared to be a really cool thing. With it I could easily make "input"
and "textarea" form elements to appear with the same size. Also links of
various colors that change color when you move a mouse pointer over them
look cute. So, if among dear readers there still are some ignorant
dumbasses like me :), the w3c specification on CSS can be found here
A couple of days ago centericq
released. It had support for external event handlers, command line
parameters to add events to outgoing queue, resize in xterm segfaults
finally fixed along with dramatically improved support for xterms in
By now, some minor problems are already discovered in the fresh release.
And I find it amusing. Why? Because before a release only obvious bugs
are usually found, despite the fact there are courious ones that daily
check out stuff from CVS, install and test it, use newly added features,
and discuss them in the mailing list
. Though to
find less obvious issues only testing by huge amount works. Which is
usually making a release in fact. By the way, the centericq related
mailing list becomes more and more nice place. For example, one guy from
Canada proposed everybody to meet up somewhere in Europe (decided it'll
be Romania) to have some beer :) The list atmosphere is extremely
friendly. Novice users get prompt explanations from guys that call
theirselves the centericq.de
etc. Looks like they're as enthusiastic about the program, as I am.
Since we're already talking about centericq, there is a need to tell
something about implementation status of such a wanted feature as the
AIM protocol support. AIM is another IM protocol from America On-line. I
said "another" because the same guys own ICQ nowdays. But despite of
this fact the ways how they treat each of the protocols look different.
Actually, AOL's treatment of open source implementations' developments
of their protocols differs. I wanted to take one of existing libraries
to make centericq support AIM which is wanted that much. Having
performed a search for such an information, I found an article (in
theregister, as far as I remember) saying about the "cat and mouse play"
between AOL and open source developers. How comes? It's easy. AOL is
trying hard to block access from unofficial client programs, and their
last move was really excellent. Here I recommend to read attentively to
understand what exactly they'd done. The protocol is logically made up
of packets, and each of them contains a number. In incoming packets a
client receives from the server there is a number which should be
answered by client with another number. Otherwise the protocol is
interrupted and the instant messaging is no longer possible, as least
for a current session. Having done a research on this matter the gaim
developer said that the number received from the server is an offset in
aim.exe - the AIM's official client executable file (sic!).
Respectively, a client has to answer with an info block taken from the
file at the offset given. Effectively, it guarantees that each AIM
service user now has to have the executable of the official client
locally. The developer of libfaim, the most wide-spread AIM
implementation of AIM OSCAR protocol in C language already gave up
Anyway, AOL's politic looks more than clear, having in mind what AOL did
to make users switch to the newest version of ICQ software. Obviously
the problems we had with the previous version of the protocol were added
intentionally. Thus, now I doubt if it's worth giving a shot.