Funny thing is … I know of so few people who actually read feeds. Even techies, many just don’t.  Even with wonderful tools like Google Reader I couldn’t persuade a smart friend to not just “reload the blog page”. Daft but true, and he picks what he does, not I. I fear that feeds are a great (simple) technology that is just out of sync with humans.

It dawned on me that Twitter’s growing success is a reaction to that problem. E-mail is boring/spamful/passe, Feeds are also a chore to keep up with, Tweets hold that promise of unsolicited  serendipity w/o the guilt. Given we don’t know we need the information in tweets we can let it wash over us, flow onto the floor when we are sleeping, and we don’t have to keep up. We can feel wired in w/o feeling the stress of something piling up.

The other aspect is the timeliness of the information. No point reading a blog posting tonight (via feed reader) that somebody was in town this morning looking for a running partner. That tweets can follow you to your phone (in near real-time) keeps folks in the loop.

Much as it goes against productivity (when at a desk) we’ll take the interruptions for those benefits.

Anyway, in order to let Wild Observers participate, I’ve wired WildObs into Twitter.

http://www.google.com/advanced_search?hl=en

I could use  “past 3-5 years” some times, but I’ll take “past year”.

It struck me that I’d enjoy following the progress of some of these job seekers, and of some of these start-ups. Amazing how quickly an association can be formed.

I didn’t take business cards, I didn’t take a mobile phone to “beam” v-cards, and I saw little of this occurring at either end of the spectrum (other than recruiters giving cards to programmers.) So, how do on-going contacts get formed after such an event? I wondered (out loud to Jeremy) if some mailing list or group w/ URLs & photos could be created. Eventually it dawned on me it already was…

Some simple “social network forensics”. Go query Twitter (the best source) and then Google:

http://search.twitter.com/search?q=boulder.me

http://www.google.com/search?sa=N&tab=nw&q=boulder.me

I quickly found most of the people I was interested in following, and a few snaps for reminders:

http://search.twitter.com/search?q=boulder.me+bkite

http://brightkite.com/objects/f3d608f2a6e111ddb9ad003048c10834

BTW: This goes to show the power of picking a simple domain name, that is itself a simple keyword. Folks didn’t need to be prompted to attach the characters “boulder.me”, they just naturally did so.

So … much as I was determined NOT to duplicate photo sharing sites (for the effort, the redundancy, and the cost) I’ve created an S3 bucket for wildobs. Today I’ll wire up file upload to a temporary directory on WildObs, and (for now) manually upload, but later change to use AWS::S3. We’ll see if “later” becomes hours or days.

One aspect of how I am working with images on WildObs is experimental, but feels good. The system stores a URL to the image, but upload creation the code pulls the image into memory to validate the URL, validate the mime type,  and attempt to extract width/height information. [One day I’ll peek for GPS information.] Knowing the width/height allows for selective resizing upon rendering, so can use one image in many context (thumbnail, medium, full-sized.)  Given improving networks and caching does this single download for multiple uses cost or save bandwidth, cost or save page load time? Dunno, I suspect cost in both but we’ll see. It sure is simpler to code. I guess the S3 bill will tell me something.

First important step for an S3 bucket … added a “please Mr robot,  don’t make me pay for your crawling”.  Let’s see if they listen.   http://wildobs.s3.amazonaws.com/robots.txt

# Robots please keep away

User-agent: *
Disallow: /

BTW: I must give a nod to this S3 Browser for OS X. Works very well, thanks!

One import part was to add the “priority” information to the module, otherwise it didn’t seem to kick in before the Phorum auth & hence failed to operate.In my case I allowed “admin” to be a local Phorum account.

<?php

/* phorum module info
title: WildObs
desc: WildObs tweaks...
version: 1.0.1
release_date: November  5th, 2008
url: http://www.wildobs.com

hook: user_authenticate|phorum_mod_wildobs_user_authenticate
priority: run module before *
*/

function phorum_mod_wildobs_user_authenticate($auth)
{
    // Only trust one admin...
    if ($auth["type"] == PHORUM_ADMIN_SESSION) {
        if ($auth["username"] == "admin") {
		$auth["user_id"] = NULL;
		return $auth;
	}
	else {
            $auth["user_id"] = FALSE;
            return $auth;
        }
    }

    // Authenticate other logins against an restful_auth DB
    $user_id = NULL; 

	// Hit the RoR database to go get the user...
 	$conn =	mysql_connect('localhost','username','pasword');
	mysql_select_db('ror_database');
	$query = sprintf("SELECT * FROM users " .
		" WHERE (login='%s' or email='%s') AND state='active'",
		mysql_real_escape_string($auth["username"]),
		mysql_real_escape_string($auth["username"]));
	$result = mysql_query($query);
	// If found, match the crypted password
	while ($row = mysql_fetch_array($result)){
		if (sha1(sprintf("--%s--%s--", $row["salt"], $auth["password"])) == $row["crypted_password"]) {
			// Keep the phorum user record synchronized with the RoR record, e.g. e-mail changes.
        		$user_id = phorum_api_user_search("username", $row["login"]);
			$user = array(
		 		"user_id"   => $user_id,
		 		"username"  => $row["login"],
		               	"real_name" => $row["fullname"],
		               	"display_name" => $row["fullname"],
				"password"  => 'not_used_here',
		 		"email"     => $row["email"],
		 		"admin"     => 0,
		 		"active"    => PHORUM_USER_ACTIVE
		 		);
		       	$user_id = phorum_api_user_save($user);
		}
	}
	mysql_close($conn);

    $auth["user_id"] = empty($user_id) ? FALSE : $user_id;
    return $auth;
}

?>

Folks talk about firefighters liking to burn things, liking fire (perhaps a little too much.) I don’t think the excitement of ignitions is that, so much as the intensity of the experience. You could be in trouble if you stop, if you slow too much, if your crew mates get out of sync, if the wind changes and the smoke and fire. In short, be in the now and don’t mess up. It is exhilarating work.

The folks in charge of this burn (as well as the other firefighters) were well prepared, cautious, experienced and safe. Everything you could want from folks leading and participating in such an operation.

I don’t know wether to mention this to all the folks manually trying to link spam, ‘cos perhaps there is some reverse spam occurring to them as they sign-up (with e-mail address) all over the internet to post a comment that gets them little or nothing. There seems some ironic justice in that.

I am looking for example/other YM4R-base sites that I can check out. Please let me know if you have knowledge of any.

I’ve found the Brightkite example, and looking at the code inside that people it sure looks like they attempt to present controls/gave informational markers. Theirs is not working either.

Ok, time to back off and see if this magically resolves itself, or if a solution comes to light. Any pointers welcomed.

That lead me to make this suggestion; i,e, set robot’s noindex until a question is marked as satisfactorily answered. Keep the robots from indexing something that is not ready.

That got rejected since /unanswered is excluded in robots.txt, but I feel the bar of a single upvote is too low.