I really hate autoconf. Its cryptic, it is a pain in the ass to use, and the only thing its got going for it is that the open source movement uses it like a bunch of lemmings.
I have went to the trouble before to make use of it. I found nothing of value from the experience other then a reminder of what it was like to write code ten or so years ago. Autoconf requires too many steps, its maintenance is high, and finding rule-sets that you need normally involves finding another project (yes I am aware of the archive) and looking to see how they did it. Who really wants to deal with M4 anymore?
What I have in mind is something a bit different. Lets put rule-sets in CPAN (yeah, I know, Perl, boo-hiss, whatever) and create an interface that builds your configure script by asking you a few questions. It should be able to pull whatever modules it needs from CPAN and then bundle them with your application so that it is portable. Have the configure script then generate make files.
Just imagine running a configure script where you will always see the same options. Perl is everywhere. Java may have Sun's marketing behind it, but Perl is the one language that is everywhere and comes on just about everything (and even Microsoft doesn't seem to have a problem with it, I have MS employees in my Perl classes at the University of Washington all of the time).
Autoconf has got to go, its time was nearly a decade ago. There is no reason to keep ourselves tied into this piece of legacy.

Isn't he just the cutest little thing?

A friend got me to thinking about "What if we just started marking out In God We Trust on all of the bills that pass through our hands?

Interesting idea... (yeah, I know, federal crime...)

Pulled this from an artile I just read:
The supersize version of the Big Mac Value Meal, for instance, which adds extra fries and a 42-ounce Coke to the sandwich, packs 1,610 calories and 63 fat grams -- more than half of even the most active person's recommended fat and caloric intake for the day.
Not that I would eat the crap in the first place, but wow. That is amazing. Gross.

I just watched Slashdot's loads cut in half after compiling lingerd. No kidding, we are well below half the load we had before across the servers. It is also funky to watch the pages load all at once. The one thing I still need to do at some point is setup a temp file system for mod_gzip to write to.
On a side note, I really wish more of the RSS grabbers and IE for the Mac would do compressed connections. It would be really nice for bandwidth.

Two plus years ago I put in some code trying to decipher what exactly was and was not a portald block in slash. See, the problem is that blocks where used as a dumping ground for pretty much anything and everything. At this point they really are not used any more except as a place to keep RSS feeds.
But the code lives on.
Right now I am trying to figure out if I was very clever two years ago and made is so you could have a static block that was an RSS feed, or if I was just being really dump and not thinking through the fact that type should be linked to the boolean in the table.
Oh, and I should kick my own ass for not writing down more notes about this then (even though I actually wrote quite a few).

Before the year is done, I will rewrite this code.

Despite the fact that we do not really point out the journal entries of users, the system is being used and after poking around for a minute I actually found a couple of journals that were somewhat interesting.
The thing is, the numbers seem to neither increase nor decrease, they just stay the same....

Me : Brian, go.
AT&T : Hello, this AT&T calling for Mr. or Ms. Aker.
Me : You have them.
AT&T : We called to thank you for being a digital cable subscriber and...
Me : I am ditching it next week.
AT&T : I am sorry, you are?
Me : Ripping the box, and mailing it back to you in small pieces.
AT&T : I do believe we have a policy against that.
Me : Fine, I will feed it to squirrels
AT&T : Well tonight I am calling you to inform you that we now have broadband access your neighborhood.
Me : You mean like a bigger cable?
AT&T : Yes, well, sort of. This will allow you to have broadband access from your home.
Me : Broad band? Like the elastic on my under-ware?
AT&T : No, this is Internet access, and for only $19.95 for the first six months we can provide you all the Internet has to offer.
Me : Isn't that rather large?
AT&T : No, you access this through your computer.
Me : You said it was a larger cable.
AT&T : No, this will come through the existing cable in your home.
Me : Then I already have it?
AT&T : No, we have to provide you with a cable modem.
Me : And I hook this up to the cable?
AT&T : Yes, we will have someone come out and install it.
Me : I have a Commodore 64.
AT&T : You run windows on that right?
Me : Sure. You can hear me typing right now.
AT&T : Then you are fine.
Me : By the way, Microsoft is loosing their hold on that trademark so you should not assume that windows means Microsoft Windows.
AT&T : I am sorry, what did you say?
Me : Don't worry about it.
Me : If I already have a modem, a computer, and a cable from you, why do I need yours?
AT&T : This is a special cable modem.
Me : Mine uses a cable.
AT&T : This is a cable modem and is not like the one you have.
Me : I have a modem, you didn't ask me what kind it is, so how do you know I don't have the right modem?
AT&T : This is a special modem.
Me : Oh, that kind of special modem. Is this the modem that I can use to find nudes of Britney Spears on the net?
AT&T : {pauses}
AT&T : It is up to you what you find on the Internet.
Me : Can I have your phone number, I think you would be cool to date.
AT&T : {click}

I just figured out last night that Jonathan Clark in Smallville, is really Bo Duke.
Yes, the music playing in his truck was the key.
And they did flip that car.

I really like Lex Luthor.

Is it just me or am I seeing an awful lot of companies masquerading crippleware as Open Source? By this I mean they release just so much of the source and hold back the rest. Anything involving performance/easier use seems to get ripped out of the open source versions and then only made available in the comercial version (which I question how legal this is).

My hope is that this bites the ass of a number of companies.

Since I originally designed the Journal system with the idea that people would read entries in mass (aka more then one at a time) I really don't have a clue as to how often people are reading each other's journals.
I know for instance who has the most friends (guess, its not hard) but I can not actually tell how many times a given entry has been read.
Maybe a general counter for just viewing someone's page?
I need to put more thought into this.

So after three years I finally replaced my Pentium II self built computer. Bigest mistake I made with it... I bought a dual proc motherboard not knowing that you needed matching pairs to make it work.
From the point of purchase till this day I doubt the existance of a matching 266 CPU for my poor computer. The box has now been put into service doing database backups (yes, I do have MySQL replication working) and doing Slash dev.
So the new computer is a dual proc 866 coppermine with a shitload of memory (its a lot) with a dual head (the card actually came from the old self built...).
I finally have a computer fast enough to run Mozilla. Mozilla i a resource hog and tends to be quite the bloatware. Thanks though to a rediculous amount of hardware I can now run it.

The scary thing?
I had X on 386BSD running in 1993 on 8 megs of RAM and it did a fine job.

Watch the bloatware go....

So, want XML/RSS instead of HTML? A number of URLs you can now attach content_type=rss to the URL and you will get XML/RSS instead of HTML. This does not work for index.pl, comments.pl and article.pl. In the future? Maybe, but for the moment no. If a slash site has an open submission bin (which slashdot does not) you can always keep track of the submissions via RSS by doing this:
http://site.com/submit.pl?op=list&content_type=rss
Long term we will probably add more formats. Its mainly a feature of extending the Slash::XML library.
Now, just to get search back online....

So, to make navigation a bit easier I added the username and my methods. For those of a technical bent, when you do a /~username/ or a /my/ you are not actually peaking into directories, instead I wrote a handler to direct you to the correct URL. Its neat and its something only perl and C can do inside of Apache.
So what are the different things you can do?

• ~username/ takes you to the user's info page
• ~username/journal/ takes you to someone's journal.
• ~username/discussions/ takes you to a list of discussions the users have created (do you know how many people don't get that they can create their own trolltalk via this??? Heck, I even added a form for creating discussions

In addition I added a feature called /my/ (you can also do just a /~/). This allows you to get at a lot of your own information. Here are a few you can do:

• /my/ this will get to your user information
• /my/discussions/ takes you to where you discussions are
• /my/journal/ takes you to where you can manage your own discussions
• /my/inbox/ takes you to any messages you have waiting. You can also use the word "messages" for inbox.

Over time we will probably add more (especially as we bring more features online). There are now quite a few hooks in the code that we have yet to exploit.
For my next trick perhaps I will document all of the places in the code you can have it spit XML instead of HTML out at you :)

So, lets look at all of our new features:
1) Fulltext -working fine
2) Innodb -working fine
3) Perl -no problems
4) Apache -works fine
5) Hardware -sucks rocks
So, back on a different DB box with replication running for backup. Lets see how long this lasts.