Permalink | Leave a comment  »

Permalink | Leave a comment  »

I'm super stoked to announce that our new website is up!

http://refynr.com

We've changed Refynr to focus on providing custom filtered Twitter feeds for conferences and events. We already have 2 conferences that have signed up, and more on the way. This type of branded dashboard is a big need, and so we hope to become a go-to source for such services. Here's our first custom dashboard - what do you think?

http://refynr.com/ocfs/

Permalink | Leave a comment  »

OpenCF Summit is so inexpensive, it doesn't seem fair. Here's the pricing, hotel and transportation details.

The list of CFML speakers includes Gert Franz (Railo), Peter J. Farrel (Mach-II), Alan Williamson (OpenBD), Matt Woodward, and many more...

It's not too big, so you won't miss one session because you chose a different one. Here's the 3-day schedule. I'm especially looking forward to the Railo training, Unconference, hanging with all the cool kidz, and learning what it means exactly to partake in cf_merriment.

I went last year, and had a blast! See you there...?

Permalink | Leave a comment  »

Hey ColdFusion friends,

Please help me out and vote for me to meet Kevin Rose (Digg, Milk, Oink, etc.), who can help tremendously in my quest to build a successful ColdFusion-powered startup - Refynr.

[UPDATE]
Note that the top 5 vote-getters win the meeting with Kevin, so I don't have to be #1 to win this amazing chance of a lifetime.

http://commonred.com/kevin-rose (vote up on the red arrow below my face [Aaron Longnion]. You may need to sign up for CommonRed)

Thanks so much!

Aaron Longnion
Founder & CTO
http://refynr.com

Permalink | Leave a comment  »

It's on Google Plus: https://plus.google.com/102177534078850027729/posts/JxuP7PVYSyi

What do you think about this serialized book being written on Google+?

Permalink | Leave a comment  »

Permalink | Leave a comment  »

I'm pumped up!

Adobe offered me a 6-month contract (with possible hire or extension), and it was too good to refuse. I will be joining a team of veteran Adobe ColdFusion developers, and will get to work on the high-traffic downloads area of the site: get.adobe.com. I haven't worked on systems with so many millions of visitors per month before, so it will be a new challenge for me.

We may also work on some AIR-based apps and advanced javascript / jQuery stuff, so that should be fun, too!

I will move to San Jose in about a week, my family will stay in Austin for a while, and we'll decide later if they should move with me or not...

What does this mean for Refynr.com, my startup that _seems_ to be doing well, but isn't actually making money?

It runs pretty well as-is, but with some minor updates it should run better, even with only minimal time from me. I can't spend much time on it for a while, but may be able to spend more money on it. I don't know the future of Refynr yet; only time will tell... ;-)

What do you think of my latest move?

Permalink | Leave a comment  »

After a few days of creating Circles, I realized that while Google+ was doing a good job of finding people from my GMail contacts that I'd want to connect with, G+ had no idea who my Twitter Followers and Friends were; I searched high and low and couldn't find a tool to find Google+ contacts from my Twitter contacts. Yesterday, I gave up searching and wrote my own tool to do it for you.

You can use Export.ly for free to get a CSV file of your Twitter Friends or Followers, then simply upload them to this tool and you'll get a list of links to peoples' Google Profiles, which very often go to their G+ Profile pages. Follow the instructions carefully.

http://refynr.com/Find_Google_Plus_Accounts_via_Twitter_Contacts.cfm

Keep in mind it's an Alpha product, so there may be bugs, it's a bit slow, and errors are possible. Let me know what you think and how it works for you.

Permalink | Leave a comment  »

via refynr.com

Check out the latest features on the Refynr.com Features page

Permalink | Leave a comment  »

We decided on this name for this book:
"From 9-to-5 Peon to Living the Dream, A Solo Entrepreneur's Journey"

Chapter 1: Can't get no...
https://plus.google.com/102177534078850027729/posts/Jf3suGCJbyH

Chapter 2: Entrepreneurial spirit 
https://plus.google.com/102177534078850027729/posts/FQUJRt7dBAZ

Chapter 3: Quitting my Job and Moving in with the In-laws
https://plus.google.com/102177534078850027729/posts/JiCefKddqZG

Chapter 4: Deciding on a Startup Idea, and How to do it Cheaply
https://plus.google.com/102177534078850027729/posts/VjtpQLtL192

Chapter 5: Finding my Way with the Early Concept
https://plus.google.com/102177534078850027729/posts/3nDwyhjYfxX

Permalink | Leave a comment  »

Now that I've created my first startup, solo and bootstrapped in 10 months, I think I have a story worth writing about.

Follow along on Google+

Read the first book written via Google Plus

Permalink | Leave a comment  »

[UPDATE 3, on July 15, 2011]: I solved the mystery of mySQL MAX(ID) queries timing out, though I don't know the root cause. Ideas?

I wrapped the query in question with a try/catch and sent the details to myself in an email with the UserID in the subject. I quickly saw a pattern that about 10 UserIDs kept showing up. I saw in the database that those accounts were for people that hadn't used Refynr for months, so I disabled their accounts and deleted their StreamItem data. Voila! The timeout errors are gone!

But I have no idea why. Some corrupt data? Some huge IDs that the system couldn't handle? Ideas?

[/UPDATE 3]

[UPDATE 4]

NOPE. That did NOT solve the problem. A couple of days later, presumably as the data grew again, the errors came back. I tried the trick from update 3 above, but new timeout errors kept coming up.

Next, I upgraded from MySQL 5.1 to 5.5. It didn't help.

Then, I even upgraded the server from 4G to 8G of RAM. That didn't help either.

I know I have poor database design, but I think in the long run I'm going to need something more scalable anyways, so I'm working on switching this table to MongoDB. So far, so good. It hasn't been too painful so far.

[/UPDATE 4]

---------------------------------------------------------------------------------------------------------------------------------

Okay, here's the situation:

1. I have a table called StreamItems that collects tweets/posts from Twitter and Facebook
2. The IDs from Twitter and FB are large numbers like 73054697145118720 and 100000057279452_212728265405748
3. The database server is dedicated for MySQL 5.1 Community, has 4G RAM, and is Quad Core
4. I've tuned the MySQL memory settings to match (I'm pretty sure I've done a decent job here, but could be wrong)
5. I get timeout errors at times when trying to get a MAX() or doing an ORDER BY on these IDs
6. Now that there's over 5 million Items saved, it can't handle the sorting necessary, so we need a better solution

My thought is to:

1. Create a table called StreamItemMaxIDs that stores the max IDs for each Twitter/FB account
2. Instead of scanning the whole StreamItems table for MAX or sorted IDs each time I need it, simply store the MAX ID in this separate table so that the query is fast every time

[I know that a NoSQL solution *might* be better (especially long-term), but am looking for a mySQL solution at the moment.]

So, should I keep the logic of updating the MAX IDs in the application-level code, or write a mySQL trigger? (Keep in mind I'm not a DBA)

Let me know if you need more info.

[UPDATE: here's the table schema]

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

CREATE TABLE `StreamItems` (
  `StreamItemID` int(11) NOT NULL AUTO_INCREMENT,
  `UserID` int(11) NOT NULL,
  `FilterID` int(11) NOT NULL DEFAULT '0',
  `ItemText` varchar(2000) NOT NULL,
  `Source` varchar(64) NOT NULL,
  `DateTimeCreated` datetime NOT NULL,
  `DateTimeLastUpdated` datetime NOT NULL,
  `ExternalItemID` varchar(40) NOT NULL,
  `ExternalObjectType` varchar(16) DEFAULT NULL,
  `ExternalDescription` varchar(2000) DEFAULT NULL,
  `ExternalCreatedDate` datetime NOT NULL,
  `ExternalSource` varchar(512) NOT NULL,
  `ExternalFavorited` bit(1) NOT NULL,
  `ExternalLikesNum` smallint(6) DEFAULT NULL,
  `ExternalLinkURL` varchar(1024) DEFAULT NULL,
  `ExternalLinkName` varchar(128) DEFAULT NULL,
  `ExternalPictureURL` varchar(1024) DEFAULT NULL,
  `ExternalReplyStatusID` varchar(64) DEFAULT NULL,
  `ExternalReplyUserID` varchar(64) DEFAULT NULL,
  `ExternalReplyScreenName` varchar(32) DEFAULT NULL,
  `ExternalUserName` varchar(64) NOT NULL,
  `ExternalUserID` bigint(20) NOT NULL,
  `ExternalUserScreenName` varchar(32) NOT NULL,
  `ExternalUserImageURL` varchar(512) NOT NULL,
  `ExternalUserURL` varchar(512) NOT NULL,
  `ExternalUserProtected` bit(1) NOT NULL,
  `ExternalShareUserID` bigint(20) DEFAULT NULL,
  `ExternalShareScreenName` varchar(32) DEFAULT NULL,
  `ExternalIsMention` bit(1) NOT NULL DEFAULT b'0',
  `ExternalIsDM` bit(1) NOT NULL DEFAULT b'0',
  PRIMARY KEY (`StreamItemID`),
  UNIQUE KEY `idx_UserID_FilterID_Source_ExternalItemID` (`UserID`,`FilterID`,`Source`,`ExternalItemID`),
  KEY `idx_ExternalCreatedDate` (`ExternalCreatedDate`),
  KEY `idx_ExternalItemID` (`ExternalItemID`),
  KEY `idx_UserID_FilterID_ExternalIsDM` (`UserID`,`FilterID`,`ExternalIsDM`),
  KEY `idx_UserID_FilterID_ExternalIsMention` (`UserID`,`FilterID`,`ExternalIsMention`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8;

[UPDATE 2: here's the dynamic SQL that times out under load]

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

<cfquery name="local.rsGetMaxExternalItemID"
	datasource="#variables.settings.DSN_master#"
	timeout="#variables.settings.databaseTimeout#">
<!--- below are samples of old MAX() statements that ran slower than ORDER BY --->
<!--- SELECT		MAX( REPLACE(ExternalItemID, '_', '') ) AS ExternalItemID --->
<!--- SELECT 		MAX( CAST(ExternalItemID AS UNSIGNED) ) AS ExternalItemID --->
SELECT		ExternalItemID
FROM		StreamItems
WHERE		UserID = <cfqueryparam value="#arguments.UserID#" cfsqltype="cf_sql_integer" />
AND			FilterID = <cfqueryparam value="#arguments.FilterID#" cfsqltype="cf_sql_integer" />
AND			Source = <cfqueryparam value="#arguments.Source#" cfsqltype="cf_sql_varchar" />
<cfif StructKeyExists(arguments, "ExternalIsMention")>
	AND		ExternalIsMention = <cfqueryparam value="#arguments.ExternalIsMention#" cfsqltype="cf_sql_bit" />
</cfif>
<cfif StructKeyExists(arguments, "ExternalIsDM")>
	AND		ExternalIsDM = <cfqueryparam value="#arguments.ExternalIsDM#" cfsqltype="cf_sql_bit" />
</cfif>
ORDER BY	ExternalCreatedDate DESC
LIMIT 1
</cfquery>

Permalink | Leave a comment  »

Okay, I know you can already use the Google Chrome Refynr app as a near-native desktop app.

So, why put it in Adobe AIR?

1. Cuz it can be an application within your PC or Mac, which makes it easier to navigate to and use.
2. You can add features on top of it - in this case: auto-refresh
3. Steve Good (http://stevegood.org/) offered to build the AIR version for free. Awesome! Thanks Steve "significantly better than" Good!!

What does it look like on my Mac?

Anyone else interested in this? We have a few small bugs to work out, and then the first Beta of Refynr Desktop may be available to test soon!

Permalink | Leave a comment  »

About 3 months back, I really wanted more feedback about my home page and web app. I knew the home pages and app needed significant improvement in the UI and User Experience.

But I felt that I had asked too much of my 500+ Beta testers already, and needed some fresh perspectives. So, when I came across http://StartupLift.com, I gave it a try. It doesn't cost much :)

So, here's the page where new users gave public feedback via StartupLift on Refynr back in March.

It's hard to be publicly "roasted", but if you stay open minded, don't defend too much, you can learn a lot! I ended up making many of the changes suggested, and the Bounce Rate (as measured by Google Analytics) has gone from 65+% to about 45% now. That's a huge improvement, and I may do it again at some point to try to improve the Bounce Rate another 10-15 points.

Permalink | Leave a comment  »

(full disclosure: I'm honored to be the brother of GEO's Founder, JD Longnion)

If you're in Austin, TX get a highly discounted oil change at one of 10 Austin locations, support a local startup, and make our world a little "greener"

Green Earth Oil Groupon Deal of the Day

Permalink | Leave a comment  »

Permalink | Leave a comment  »

A new Refynr customer sent me an email pointing to this great TED talk warning to beware of "filter bubbles", and the user wrote, "I believe you need to do without the filters. I'm not going to be talking about something I don't know about and so some relevant and important tweets will be screened out. ... This TED talk explains why filters screen out the kind of thing that while not of immediate interest - are relevant.", and I replied to defend Refynr:

I love that TED talk, and have tweeted it and shared it with many friends... but I actually think the "filter bubbles" he's referring to are the auto-magic, algorithmic ones like Google search, FB News Feeds, and even smartr. Those tools are only *guessing* at what you want to see. Refynr puts you in control. If one week, you want to filter out everything about the "royal wedding" or "Opra", you can do that. If one day, you want to make sure you don't miss anything about a certain football match, add that to your Include filters. If you want to mute a certain @user or #stupidConference, just add it to your Exclusions ;-)  Some people don't want filters at all, some want an algorithm to do the filtering for them, but I want to be able to explicitly filter whatever I want at any given time - so that's what I created Refynr.

Another friend on Facebook sent me the following question, and also referred me to the TED talk: "Is Refynr a bad idea?" I replied:

that's a great TED talk. I've seen it and have been thinking about it a fair amount... 

I think the talk is more referring to Google, FB, and other services that are auto-curating content without your permission, thus removing the control from you. That's the biggest problem, but there's also the problem where people are putting their head in the sand, creating those "filter bubbles" for themselves.

Refynr, or similar explicit filter systems, I believe are essential for people or businesses who are very active on Twitter... once you follow more than 50-100 people, the information overload becomes just too much. People can go too far, and filter out everything except that which is in their comfort zone. It's my hope that people will not filter *that* much, but will create filters in such a way that they still learn new things, especially information that is important for people to know about the world.

Too information is a bad idea.
Over-filtering is a bad idea.
Auto-filtering, without consent, is a bad idea.

For the most part, people use Refynr to intelligently filter the info overload, so I think it's a good idea.

So, now I want to know: what do you think?

Permalink | Leave a comment  »

Since I'm not that great with web design, CSS, or UX I've started simple with the design of Refynr.com's web app. This is a bad thing at times for my users, who expect a better user experience, so I incrementally improve over time. 

One problem was that the content is made up of streams/feeds of data from Twitter & Facebook, so there was a lot of scrolling. As you scrolled in the old version, you lost quick access to the navigation. Being totally out of touch with CSS these days, I didn't even know that I could have a "fixed" header with pure CSS; I actually thought I would have to use dreaded frames! And I had also investigated doing a fancy jQuery menu that annoyingly followed you down the screen, complete with bouncy animations... glad I didn't go that route!

I noticed that my blog theme had a fixed menu (see, on the left), and viewed source: what?! No frames?! So, I did some googling and found a great, beginner article (perfect for me!) to show my how to create fixed headers (or footers) via CSS: http://www.noobcube.com/tutorials/html-css/fixed-header-footer-layout-a-begin...

It was pretty easy, I didn't have to reoganize my HTML markup (except for adding an id or two to some DIVs), and voila, it works! I tested on Firefox 3.6, Chrome, Safari, and IE 7, 8 & 9... guess this technique has been available for a while... *sigh*. 

Here's what the old version looked like:

But here's the new version... w00t!

See the full gallery on Posterous

Permalink | Leave a comment  »

Permalink | Leave a comment  »