Cool idea amelinda, cool contest idea jc. ++'s all around.

vroom | Tim Vroom | vroom@cs.hope.edu

You'll have to put a moderator or two in charge and eventually PM should hold the pre-bought prizes till the judgement comes in. Once a script wins, we pick it over for safety and power-power-power as group thanks to the sponsor.

Or not... but in any case Go Go jcwren!

--
$you = new YOU;
honk() if $you->love(perl)

> 7. Entries must run under Perl 5.005_05 and below.

Er, how far below?? There comes a point where punch cards come into play, and you can't email them as a .tar

A couple of carifications:

Perl versions:

Basically, an entry cannot be reliant on any feature of 5.6 or above, but must use a Perl 5 version.

Windows / Unix compatibility:

I realize that not everyone will have a Windows box to test on, and others may not have a Unix box. However, I feel this is a pretty generic challenge. I believe that if fork() is avoided, there shouldn't be any OS specific functionality. If you select a CPAN module for your DB (and this is not REQUIRED. The requirement is only that headlines be stored. You could use a flat file.), the module generally indicates what platforms are or are not supported.

Meta tag refresh time:

The parameter for the META tag refresh interval is in minutes, with a default value of 10. (In the rules I had not clarified what units the refresh interval should be expressed in.)

It's a little more difficult to design a challenge that gives all platform users a fair chance. This should run on a Mac, although so few people use Macs (compared to Windows/Unix), that I did not specifically call them out. Since this is the first one, I'm open to suggestions for future challanges on how to handle this. But for now, it was as fair as I could think of. Any input is appreciated.

If I've still not expressed this well, please /msg me and I will update this node to clarify.

The www.tinymicros.com/contest page has been updated, and includes a list of changes.

--Chris

e-mail jcwren

Any Perl 5 version? I think many here would have trouble testing that. For instance how many here know (or need to) that

foreach my $foo (@bar) {
  # Do something
}
[download]

is a syntax error in 5.003?

Did I miss a memo? I don't recall seeing anything on who won this.

Cheers,
KM

update it doesn't look like I am, since this is the file that the ticker points to... but if you wait long enough, you'll see what looks like an error in their regexps whenever one of the state headlines has an '&' in it. the headline gets truncated at the '&'!

update Il Papa JCWren has kindly contacted the CNN folks about the stale news & apparently they've had a hung ftp server for a month+. Wow.

update It looks like somethings gone all b0rken again: headlinecount=0&&headlineload=done

"There is no news. Go away."

No, you're not being dense. I used tcpdump to get the URL that the Super Ticker uses to display it. This is indeed the file it's using, but it *is* out of date.

I have a call in to CNN right now trying to figure out what their problem is.

In the event this is not fixed by Friday afternoon, I will provide an alternate URL, and write something that generates random headlines that will simulate that file.

Since CNN advertises the Super Ticker regularly on Headline News, I suspect they will be interested in getting that fixed.

--Chris

e-mail jcwren

Maybe a split level contest where code submitted by Abbot's and up are judged separatly from code submitted by lower ranking monks.

Some of us are only lower in rank because we've not been here for that long.

Beware the Initiates!

--

Brother Marvell

--
Ryan Koppenhaver, Aspiring Perl Hacker
"I ask for so little. Just fear me, love me, do as I say and I will be your slave."

As new headlines come in, they are placed in the database. Headlines older than 24 hours (by default. This is command line controllable) should be aged out. All headlines currently in the file should be displayed.

Think of it as a sliding window, with a window width controlled by the command line argument for aging.

--Chris

e-mail jcwren

Thanks. I would have figured as much, but with that being the case, I can't figure any interpretation of...

• ...a list of all states that headlines have been encountered for...
• (states with no headlines should not have an active link)

...wherein the second requirement is meaningful. With only one set of headlines (as opposed to a "new from the web" and an "old in the file" set), the set of states for which there is a headline, and for which there is no headline is null, thus all states listed will have an active link.

Or am I missing something?

--
Ryan Koppenhaver, Aspiring Perl Hacker
"I ask for so little. Just fear me, love me, do as I say and I will be your slave."

-- Randal L. Schwartz, Perl hacker

AgentM Systems nor Nasca Enterprises nor Bone::Easy nor Macperl is responsible for the comments made by AgentM. Remember, you can build any logical system with NOR.

I gave that issue some consideration, and decided that it raised more problems than it solved. Different people may take fairly radically different approaches to solve the problem.

It can very difficult to make a judgement call on someone elses code. What I consider elegant may not be what you, or anyone else, does. I felt if I was the sole judge, I might appear arbitrary as to why I selected one and not another. In addition, getting multiple people involved in judging involves time commitments that I have no way of managing. And to be honest, I didn't feel like fielding calls of "Well, this way is better, why didn't you pick that"?

I also thought about saying "Please don't start coding until 2000/12/01", but knew that I'd have no control over that, unless I with held a critical piece of information until that date. If this turns out to be a problem, maybe next time I'll do it that way.

I'd like to do this every month. Maybe the first one won't go off without a hitch. That's why I'm asking for input, and taking feedback. I see one occurrence already of where I failed to clarify the specifications enough.

And the very real reality is that no contest is completely fair. You won't win the $1,000,000 from McDonalds if you can't go there to get game pieces, and you have no stamps to mail in for pieces with. I feel like I've tried to be as fair as possible, within reason. If someone disagrees, I'll factor that in. But I seriously doubt that no matter how it's done, everyone will be happy.

--Chris

e-mail jcwren

Perhaps, then, you can offer categories (obviously not passing out T-shirts to everyone...):
• Most Obfuscated Solution
• Shortest Solution (character-wise)
• Most Ridiculous Solution (using the wrong modules for the wrong job but results in correct solution!)
• Most Wasteful Solution (without dumping core)
• Most Clear Solution (without documentation)
• Fastest Solution (via BenchMark assuming taking web page from file instead of net)
• perhaps some others...
Arguably, not all of these are interesting or useful and are, of course, at your discretion, but I think it would be interesting to give lots of people chances to get something interesting done for others to see. Perhaps it would be more fair for all of the applicants' progs to show up on tinymicros.com for all the perlmonks to vote on (anonymous programs until winner declared for fairness). That way, the system can be almost completely automated.


AgentM Systems nor Nasca Enterprises nor Bone::Easy nor Macperl is responsible for the comments made by AgentM. Remember, you can build any logical system with NOR.

"...command line switch that allows font name specification for the text, and uses them correctly for all text on the page..."

Use the CSS, Luke! I hope the readership isn't thinking of using the FONT tag; that's a "Bad Idea".

Some may consider the use of a FONT tag gauche. However, since it's a command line argument, and therefore optional, and a heck a lot easier for most people that CSS sheets (which have their own problems), the FONT tag is completely acceptable for this challenge.

--Chris

e-mail jcwren