Oracle UCM 11g Now Released!

Good news! The 11g version of Oracle UCM is finally available! This version is a bit of a re-write to run on top of the WebLogic application server. Oracle has been talking about this release for some time, so I'm glad to see it finally available.

Despite the architecture change, the basic administration interface should be familiar to existing UCM customers. The install is a lot different, and you'll need to learn basic WebLogic application server administration. This isn't as tough as it sounds... you can refer to my presentation on the top 10 things UCM admin need to know about WebLogic to get started. Once you wrap your head around the concepts of instances, servers, and domains, it's not too difficult.

There are a ton of new performance enhancements in this release... when using a half-rack Exadata Machine, they were able to get 180 million 100k document check-ins per day. Note: these are full document check-ins; not merely lightweight "object" storage. Now that UCM is the repository for IPM, you can imagine how Exadata makes an excellent addition to a high-ingestion setup.

Download performance is pretty good as well... using the new Open WCM technologies, Oracle was getting about 120 Site Studio web page hits per second on commodity hardware... Is that a lot? Well, by way of comparison, Wikipedia has a farm of 249 servers, and each node gets on average 80-200 hits per second per server... sometimes spiking to 250... and Wikipedia is using a very lightweight LAMP stack, as opposed to the full enterprise JEE stack that Oracle was using for their tests. I'm sure with a additional tuning (like a CDN) you could get more than 120 hits per second with UCM... but as a general rule, if you need more than 120 hits per second, you really need more than one server!

You can download the new software in the usual place. The documentation is bundled with the rest of the Oracle Fusion Middleware documentation,

Since 11g is so new... Oracle University doesn't have any training for it yet. Luckily, my company -- Bezzotech -- can do Oracle UCM 11g training if you need it.

Oracle UCM jQuery Plugin for AJAX

I recently put together a jQuery plugin for Oracle UCM, and thought I'd share. This connector allows you to use jQuery to make UCM Service calls through AJAX, and easily display the results. This is 100% pure JavaScript, no Java, Idoc, or ADF required! You can download the plugin from the Bezzotech library page.

You will need to be familiar with jQuery before you can use this plug-in. You can start with the jQuery tutorial at W3Schools.

This plug-in is essentially a collection of wrappers around the $.post() method in jQuery, to allow you to more easily make AJAX calls to Oracle UCM. For example, in order to make a "Test Connection" button to call the "PING_SERVER" service, and pop-up a status message, you'd use code like this:

<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.4.2/jquery.min.js"></script> <script type="text/javascript" src="jquery.oracle-ucm-1.0.js"></script> <form> <button id="pingServer" name="pingServer">Test Connection</button> </form> <script> $(document).ready(function(){ $("#pingServer").click(function(){ var testData = { IdcService : "PING_SERVER" }; $.ucm.executeService(testData, function(ucmResponse){ var msg = ucmResponse.getValue("StatusMessage"); alert(msg); }); return false; }); }); </script>

Notice how we have to pass function(ucmResponse){} as a parameter to the executeService method... and then fill that function with the code we want to run after the request completes. Also note that we are passing in functions as parameters to the button's click event, and to the document's ready event. Passing around functions in this way looks quite odd if you're used to object-oriented languages like Java... but it's standard operating procedure for jQuery and AJAX.

The ucmResponse object is a standard JSON data object, with the following convenience functions added:

  • getValue('value') looks in the LocalData and the ResultSets to get a value from the response.
  • getRow('resultSetName', index) gets the row in the specific ResultSet at the specific index as an associative array (hashtable) of name-value pairs.
  • findRowIndex('resultSetName', 'column', 'value') returns the first row index in the result set where the column equals the specified value.

There are more examples and instructions in the download. If you have any suggestions or requests, please leave a comment!

What Almost Everybody Gets Wrong About Collaboration

I've used dozens of collaboration systems... none of which really stood out to me. It wasn't that they were difficult to use, it's that none of them actually solved the human problems that limit our ability -- and our desire -- to actually collaborate.

It wasn't until recently that I came across a talk from Clay Shirky, which explained pretty well what was missing... Clay spoke about human nature and software and asked a very important question: why do some kinds of sharing work well, while other fail?

Well... one reason is that according to anthropologists, there really is no one thing called "sharing." We humans -- like all primates -- have three distinct ways that we share... and our brains are wired to do different things based on what kinds of sharing we are doing.

For example... I want you to imagine that a little old lady is walking up to you on the street. She makes direct eye contact, and gestures that she has a question for you. I want you to take a deep breath and genuinely imagine that she asks you one of the following three things... and take note of your emotions:

  1. she asks you for money,
  2. she asks you to help her cross the street,
  3. she asks you for directions to the bus stop

If you are like most primates, your initial gut reaction to #1 is something like "NO! MINE!" Your gut reaction to #2 is "eh... OK..." And your gut reaction to #3 is "Absolutely! I'd be happy to!"

Why??? All three are sharing, aren't they? Not quite... millions of years of evolution have wired us to react differently to different kinds of sharing. The examples above each demonstrate one kind of sharing:

  1. Sharing Goods: the gut reaction it to feel bad when you give somebody else your goods... because then you can't use them anymore, and you might not be able to replenish them. Even generous people have this initial reaction.
  2. Sharing Services: people are more generous with favors, because they don't lose anything physical... merely their time. However, before sharing your time, everybody does a little mental math. Do I have the time? Is this worth my time, or should I delegate to somebody else? Shouldn't I be compensated for my time?
  3. Sharing Information: people are most generous when it comes to sharing information... it takes little measurable time, it costs nothing, and sharing information makes us feel good. We feel good, because we feel like we've helped out one like us, and made the world a better, more knowledgeable place.

Clay used the example of Napster to illustrate his point... it took a goods sharing problem (can I have your CD?) and a service sharing problem (can you make me a mix tape?) and turned it into an information sharing problem (can I download all your already ripped albums?). People were sharing their albums online because it made them feel good.

Like monkeys with iPods...

The problem with most collaboration software is that collaboration software relies too much on "service sharing" to get people to take action. I post some information on a place for "sharing" and to make it better through input from others... but in order for that to happen, first you need to read it and understand it. That's sometimes not a big deal, but in many cases it's a significant time investment.

To make matters worse, some of these systems even make it difficult for you to do the mental math for you to determine whether reading my document is worth your time... Is this for an important project? How important? Do you need my expertise for all of this, or just a few pages? Should I be charging your department for my time? Not only is this still a "service sharing" problem, but a pretty tough one at that...

Ideally, a good collaboration system would obey the 2 minute rule. Getting information is still something of a service... but if it's a service that can be performed in under 2 minutes, it will probably "feel" more like an information problem... which makes it more likely to be done. If it takes more than 2 minutes, then it feels like a service problem, and then we're back to the mental math problem...

Getting down to the 2 minute rule is tricky... you could opt for a system like Aardvark, which tries to match simple questions with the right person to answer it... Alternatively, you could force people to jump through a few hoops first before asking a question; essentially making it easy for people to answer your question. If people can estimate the difficulty of the task and the value provided by the solution, then it's easier for them to do the mental math for the tougher problems.

Neither of these are new concepts... in fact bug tracking systems for successful open source projects use a blend of both. They'd have to, or their entire model would collapse! Although I have yet to see any enterprise level collaboration system truly adopt these concepts... probably because the enterprise is something of a captive audience. If you're lucky youll have a system that focuses on ease-of-use and good training... but adapting to human behavior isn't always high on the list. Would people still use your collaboration system if you didn't pay them? Probably not... which usually means a problem...

Hopefully the big push to "Enterprise 2.0" solutions will get more software companies thinking about making software that's a natural extension of human behavior... Maybe in a few years we'll have Aardvark for the enterprise... but I'll take my standard curmudgeony "wait and see" attitude ;-)

In Migration Mode...

You might have noticed this site go away for a while yesterday... my ISP is migrating my site to the latest version of Drupal, and we hit a few snags when we tried to take it live...

We're giving it another whirl today, and hopefully it will be good-to-go tomorrow. I have a new skin that I'll be showing off... as well as a few new features.

The World's Oldest Computer

Sometimes people ask me, "what does your company's logo mean?" When I founded Bezzotech 4 years back, I wanted a symbol that says "high-tech and new", but also something that says "old school rocks!"

After a bit of thinking, I realized the answer was simple. There was only one appropriate symbol, the computer invented by the ancient Greeks: the astrolabe.

An astrolabe is essentially an analog calculator. You would set it according to your latitude, and point the rule at a well known object in the sky: the sun, the moon, the north start, etc. Depending on what you picked, you could calculate the time of day, the time of year, or your location. It could be used for things as simple as calculating when to plant your crops, or as complex as a geographic survey of an entire city. Advanced ones had charts on the back to help with math calculations, and conversions.

It was originally invented in Greece in approximately 150 BC, and spread quickly through Europe and the Middle East... They aren't very common today, but 2000 years ago they had approximately the "cool factor" of the iPhone. One Persian astronomer wrote a book on the 1000 different uses of an Astrolabe. Technically, the astrolabe has more apps than the Google phone...

You can make a simple astrolabe using paper or wood... but of course the most beautiful ones are made from brass. In the old days, an educated person would not only know how to use an astrolabe, but they could build one as well. This presentation from the TED conference covers how it was used, and how amazingly resourceful our predecessors were with them:

So what do you think? Does my logo look like an astrolabe, or is it just a cool gear thingie?

Blog Year In Review

As usual, I'm celebrating the end of my blog's fiscal year on April 29, 2010... because that's the day of my first blog post. Also, I didn't want my "end of year" post to be drowned out by everybody else's end of year posts...

Anyway, at the end of my blog's year, I do some analysis on what posts were the most popular. I use Google Analytics on my blog to collect this info, ever since I found other engines to not be what I needed... see my post on how many hits does your site really get for more info.

Overall, the blog is still doing well... I got 182,364 page views, which is about 15% more traffic this year than last year... below are the top posts of 2009, according to Google:

So.. although this may seem a bit early -- or a bit late -- happy new year! And hopefully you're following you new years resolutions.

The Top 10 Things Oracle UCM Users Need To Know About WebLogic

As promised, below is my presentation from IOUG Collaborate, 2010. It's an intro to WebLogic for people who know Oracle Universal Content Management:

In case you haven't heard, the 11g version of Oracle UCM will by default run on top of WebLogic. I have heard that support for other platforms are planned post-11g -- such as JBoss, WebSphere, and maybe even the old-fashioned stand-alone mode -- but it's still a good idea to become familiar with WebLogic. These concepts will be similar on JBoss and WebSphere... especially when it comes to the security and performance sections of my talk.

Enjoy!

Bezzotech and IRA Merge Into One!

This shouldn't come as a surprise to my clients, but it might to a lot of my readers...

About 6 months ago, Jason Clarkin (from Impement R Advantage) and I had discussed joining forces. We talked a lot about the kind of firm we could build together, and what kind of new software we could create. We started working very closely together to make sure our business styles were similar, and that our skills complimented each other's well...

Once comfortable with that, we decided to pull the trigger an make it official. We are now one company!

We decided to make the big splash at IOUG Collaborate... together, our company is giving three UCM related talks this year:

  • Sesion 418: The Top 10 Things Oracle UCM Customers Need To Know About WebLogic. Tuesday, April 20 3:15 p.m.
  • Session 422: UCM as the Repository, Presentation Layer of Your Choice: The Future of WCM. Wednesday, April 21 8:00 a.m.
  • Session 107: Impact of Oracle UCM on Organizations. Thursday, April 22, 11:00 a.m.

Also, be sure to swing by our booth: #1743. We're right by the big Oracle booth near a few other UCM vendors. See you there!

Twitter is Sooooooo 18th Century...

Just in case you think that Twitter is cutting edge, the good people at the Wall Street Journal are here to set you straight... On their blog they interviewed Cornell professor Lee Humphreys, who did an analysis of the striking similarities between Twitter and 18th century diaries.

Personal journals were not always about writing long prose about everything you did that day... most of them contained very short one-sentence updates:

"Before the end of the 19th century, diaries weren’t considered private or introspective. Instead, people wrote semi-public diaries that were often shared among faraway family members and others. And space was at a premium; by the mid-1800s, popular “pocket diaries” were only about 2 inches by 4 inches and were intended to be more mobile than earlier books."

Take for example somebody I just started following recently: John Quincy Adams. Yep... son of John Adams and the sixth president of the USA, John Quincy is up on Twitter. The Massachusetts Historical Society is taking his personal diary verbatim and placing it up on Twitter... 200 years to the day AFTER he originally wrote them, and most of them fit into the 140-character limit just fine! Call me nerdy, but that's pretty cool.

One of my favorite lines from the article is this:

Dr. Humphreys said the research serves as a good reminder that not everything in new media is entirely new. "It’s helpful to put things into historical context," Dr. Humphreys said. "It’s amazing how much human nature hasn’t really changed all that much."

Too true... when Twitter first came out I thought it was a silly concept... why would I Tweet when I could Blog? The point it, despite the number of blogs out there, not many people enjoy writing... but everybody like keeping up-to-date with friends.

But of course, that's what Facebook is for ;-)

Almost Time For IOUG Collaborate 2010

Anybody else jazzed that IOUG Collaborate 2010 starts in a week??? Hanging out with Oracle Nerds and talking technology may not be everybody's cup of tea... but it's waaaaaaaaaay more fun than most people realize... Honest! No foolin!

...or maybe I just have unusual tastes...

In either case, I'm going... and I'll be giving my talk on The Top 10 Things Oracle UCM Customers Need To Know About WebLogic:

Date: Tuesday, April 20, 2010
Time: 3:15:00 PM - 4:15:00 PM
Location: PALM D (wherever that is)

In case you hadn't heard, both IPM 11g and IRM 11g now run on top of the WebLogic application server. The word is that Oracle UCM 11g will need WebLogic as well... No word yet if this will be a requirement, or if some kind of "stand alone" mode will also be supported. In any event, the go-forward strategy for pretty much all Oracle middleware products will center around WebLogic... so it's probably a good idea to get to know it sooner rather than later.

I've used WebLogic for years, and recently I've been playing with a beta copy of UCM 11g on WebLogic. I thought I share my thoughts and help demystify what this WebLogic stuff means to UCM customers... Spoiler alert: everything will be just fine.

Anyway, if you want to meet up at Collaborate, let me know! I'll be attending a few of the 15 sessions on UCM -- a shockingly number, IMHO -- and attending the Oracle ACE dinner on Sunday... but my schedule is otherwise quite open...

Hope to see you there!

Let it go, this too shall pass...

I couldn't resist... this latest video from OK GO is as amazing as their other ones...

Should Product Managers Know How To Code? Steve Jobs Couldn't...

Last week I was in a spirited debate with @jkuramot over at The Oracle AppsLab about whether or not Product Managers need to know how to code.

Jake says yes, I say no... Primarily because we disagree about what a product manager actually does...

First, I think I should answer the question, what the heck does a Product Manager do all day long??? Most of the time when my wife tells people she's a "proDUCT manager" they think she's a "proJECT manager"... which isn't even close. I've met quite a number of Product Managers, in different industries, and I can safely say that most of them do significantly different tasks... And many of them disagree on what their main focus should be.

Why such contention? The best explanation I ever heard was this: a product manager is more or less the CEO of a product line. Which means that pretty much anything that will help your product, you should do... which means a million different things in a million different situations. In order to be successful, you need to know a little bit of everything -- what features to add, what new markets to attack, what sales people need to sell -- in order to be an effective Product Manager.

Rich Manalang did not like that definition, saying it was too broad... he preferred the idea that a Product Manager is responsible for just the "life cycle" of the product. In other words, vision, design, creation, testing, support, and end-of-life decisions. In my opinion, this definition is far too narrow... it's far too "developer centric" in that you focus mainly on new features, training, and support... but neglect such very important things like what effect does the existence of this new product have on the rest of the company? An individual Product Manager might not know this... but hopefully somebody on the Product Management team does! If not, then nobody is building the path between a successful product launch, and a successful company.

If the Product Manager isn't responsible for that critical task, then who the heck is???

Back to the original question... do technology product managers need to know how to code? I say, emphatically no... It's a useful skill to have, don't get me wrong... but I disagree that it's a requirement for everybody on your product management team.

Let's be clear -- programming skills are primarily useful to a Product Manager as a communication technique. A prototype speaks volumes about what features people want... but that's about the limits of it's usefulness to a Product Manager. And, of course, if you are a good communicator, you can certainly do without it.

"the three great virtues of a programmer: laziness, impatience, and hubris" -- Larry Wall, inventor of Perl

Now... what if you are a Product Manager in charge of lazy, lazy developers? This happens. Maybe you want a feature, but the developers don't want to add it. So the developers give you the run-around so they can go back to playing Halo. Well, in those cases it helps to know how to program, so you can call out your developers for being lazy virtuous... but this only works if you know a great deal of the existing code base as well! Just because it works in a prototype, that doesn't mean it will work when integrated into the product.

"When it comes to understanding code, if you wrote it 6 months ago, it might as well have been written by somebody else" -- Ancient Geek Proverb

Knowing the code base is a pretty hefty requirement... even seasoned developers don't know everything about their product... so it would be nigh impossible for a Product Manager to do so. It's more important that their minions think they know the whole code base, to try to keep the lazy virtuous developers honest. The best technical Product Managers know ho to "dive deep" into the product, and know well a handful of obscure but important details about the system... this inspires a healthy amount of fear.

Ultimately, Product Management is so important and so difficult, that it's almost impossible to find all of the skills you need in one person. Small companies do this, but as companies grow, they usually break it down into three teams... it's occasionally useful for the "technical" Product Manager to know how to code, but this rule does not apply to your whole team.

If you're looking for more info on this subject, I've heard great things about the Pragmatic Marketing Framework for designing a Product Management team.

Publishing is Dead...

Or is it?

The Origin Of Bex: Now Available On Netflix

I'm frequently asked where I got the nickname "Bex." Back in 2008 I finally put the matter to rest by explaining the origin of "Bex".

The quick story is this: a while back I saw a British TV show with a character named "Bexley," thought it was a cool name, and started using "Bex" as one of my (many) internet aliases. When I went to college, there were too many "Brians" in my dorm, so they decided I needed a nickname... one of my geekier dorm-mates asked me what I used for internet aliases, I mentioned them, and they liked "Bex" the best.

The rest, as they say, is history...

The TV show in question is Red Dwarf, which is a something of a sci-fi cult comedy... I was recently surprised to see that Red Dwarf is available via NetFlix-On-Demand! If you have NetFlix, and about 20 minutes to spare, you might want to watch the television episode that spawned my nickname. And maybe a few more episodes, if you care to...

The show is well written, if not a bit quirky... but, if you're a frequent reader of my blog, you probably have the sense of humor necessary to find it charming ;-)

Why People Just Don't Pay for News

CMS Report has an interesting rant about micropayments, and how they never got off the ground... many people have tried to convince me to pay a quarter or a nickle to view their online content, but I've never done so. Every few years, somebody comes up with some master plan based on the theory that "this time it's different!!!" But sadly -- and totally predictably -- they failed.

Why don't people pay for news? Because the most powerful word in marketing is "Free." No matter how little you charge for "quality" content, if somebody else is offering a reasonable substitute for free, you will always lose.

The latest "big idea" in this history of failures is Rupert Murdoch's attempt to charge for their online content. Some folks see Apple's new iPad as a game changer here, perhaps shaking up the market and getting people to pay for quality content. I'm skeptical... Yes, the iPad is pretty, and yes it is probably the best possible platform that "paid content" could ever hope for... but that doesn't change the economic realities.

Yes... the Wall Street Journal's articles might be exceptional... they might be light years better than what you can find for free on blogs and Bloomberg.com... but how can Murdoch prove to a skeptic that "paid-for" content is worth the extra cost? Unless they give away the whole article for free, nobody can judge it's quality. Also, just because one article was great, does not mean future articles will be great... Finally, if it really is a great article, people will blog about it, or editorialize about it, after which I can find a decent summary elsewhere.

People just don't have much brand loyalty to information sources anymore... Whoever gets it to me in the way I want it, will win my loyalty for today... but once you're boring, or ask me to login, or ask me to pay, then I might take my eyeballs to one of the other bazillion sites out there.

News is a commodity, and therefore subject to the economics of commodities. There is a little bit of profit in their creation, but much more in their distribution. In the past, the newspapers owned their distribution channel: printers, packers, delivery trucks... heck, the New York Times even owns their own forests and paper mills! The majority of their expenses are spent maintaining their distribution channel, not in paying people to write quality content.

Rupert Murdoch whines very loudly that his content is valuable and Google should have to pay to spider it... what he's really saying is that he's mad that the internet has made his distribution channels less profitable. If Fox truly cared about creating "quality content," they'd probably drop half their sitcoms.

Is there a way to save newspapers? Sort of...

Obviously, companies with good content need to get into the new distribution channels if they want to survive. The NBC/Comcast merger is a good example... although as a consumer I'm not a fan of so much power being in one single entity. I hope other companies get into the residential high-speed internet business so we have more competition... I'm happy to see that Google is getting into the residential ISP business, and I hope to see more competition soon...

In other words... The New York Times and the Wall Street Journal will survive... but their distribution channels will not. The sooner they get out of the dead-tree-scattering business, the better!

Tricksy Oracle Releases IPM 11g Early!

A lot of folks have been waiting for Oracle to release their 11g version of their ECM suite. We were told to expect it calendar year 2010... which I guess means early late February!

Universal Content Management 11g is not released yet, but Information Rights Management (IRM) and Imaging and Process Management (IPM) are bundled into the same installer, and currently available. You can download the IPM/IRM 11g bundle off of Oracle's web site.

Be warned, the download is over a gigabyte, so pack a lunch...

This is a complete rewrite of IPM so that it runs on Fusion Middleware. It uses BPEL based workflows instead of the previous custom workflow engine... it stores content in the UCM repository instead of a separate one... and it comes with tons of connectors to enterprise applications such as E-Business Suite. There are also solution templates for common business workflows that are good starting points.

If you got an hour, I'd recommend viewing the December 2009 ECM community webcast to get a brief overview of what's in IPM. Also, There will be a IPM Webcast this Thursday Feb. 25th at 11:00 a.m. PST / 2:00 p.m. EST. Oracle also has a IPM Launch Center with a few more goodies. If you got even longer, read the documentation.

I Called It! Google Acquires "Social Search" Company Aardvark

Man... I was just blogging last week about how social search systems like Aardvark would completely change the dynamic of search engines... particularly for mobile users. I said that it was perhaps the best sales model for folks like Twitter and Facebook to challenge the dominance of Google.

Well, too late boys! Because it appears Google just acquired Aardvark. I anticipate Google will be using their standard ad-engine to target Aardvark users... and probably use it to make their mobile offerings more socially aware. This will continue to cement Google's place as the #1 search engine, and gives them a pretty cool foot in the door of actually taping into the useful aspects of your network of friends.

@Twitter: next time you should be faster! Maybe if I tweeted my advice in 140 characters instead of doing a well-thought-out blog post, they would have been able to beat Google to the punch.

Apple iPad: Another Reason HTML5 will Beat Flash

Many people have been using Adobe technology for next generation "rich internet applications." Many folks -- myself included -- have warned against this kind of behavior. Flash is a stopgap solution with a lot of cool features, but it would be unwise to use it as a long-term standard. I mentioned earlier in my HTML 5 versus Flash/Flex post, that enough of these flashy features will be a part of HTML 5 and WebKit enabled browsers... so you might want to start pushing them as a requirement.

Another reason why you should avoid Flash? The new Apple iPad.

The jury is still out about this one... Steve Jobs says it's the most important accomplishment of his life, while haters are bashing it almost as much as they bashed the original iPod. Personally, I'm disappointed in its lack of a camera or voice recorder. Other than that I'll have to kick the tires on it a bit.

However, some folks are shocked that the iPad will not run Flash or Flex applications! You know all that time and effort you spent making flashy gizmos for your web app? Well, get ready to do it all over again... because the new iPad does not run Mac OS X. It runs a modified version of the iPhone OS, which never, ever, ever supported Flash... and there are zero plans to support it.

Likely, the future is going to be CSS, JS, and HTML 5. Yes, I know Flash fanboys will say that those technologies are just plain awful when you want total control over the look and feel across multiple browsers... and they are 100% correct. But, sorry to say, it's not always the best technology that wins.

UPDATE: it appears that Steve Jobs has gone on records about why the iPad and iPod will NEVER support Flash. Steve-o brings up a few more reasons I did not cover here: Flash is a power hog, it doesn't support "touch" interfaces, and it crashes a lot. Steve Jobs ends with a plea: Adobe should use its brainpower to make a cross-platform IDE for HTML5, and stop trying to cram Flash down our throats. If they don't, then the "next Adobe" certainly will...

Social Search: How Facebook And Twitter Can Conquer Google

Oreilly has an interesting article about the future of search engines...

It was prompted by the annual report by Aardvark, which is a pretty cool little startup. It's similar to Google Answers and Yahoo Answers, where you can ask any question and it will get you an answer. Sounds simple enough, but the devil's in the details. One thing that impresses me about Aardvark is that they really seem to "get it" when it comes to social search: people aren't looking for THE answer, they are looking for AN answer. They call this the "village" model, as opposed to the "library" model preferred by Google.

Aardvark uses your profile to learn a few things about you: where you live, interests, hobbies, etc., and then tries to find a similar person in their network to answer your question. It also connects with Facebook to figure out a few more things about you... who your friends are, what they like. Essentially, what your "villiage" is. Once it figures out what virtual village to put you in, it forwards your question to somebody likely to answer. If you ask a question about cameras, it forwards your question to some local camera experts who have been helpful in the past.

Sweet...

I'm impressed with the technology in general... also the participation rates are amazing. About 88% of questions sent to Aardvark got answered, and 75% of people who asked a question also answered one. Most interesting, is how mobile users were the ones asking the most questions. Sure, if they were sitting at their computer, the might fire up Google, or an IM client, or dig through their email archives... but despite all attempts to make them usable, mobile devices are still terrible at user input. Also, the

When you're out and about, and want to know where the best place to park is, how would you find that out?

If you have a smart phone, You could Google around to find a forum about the best parking spots in your specific city. Then scour through the archives, and find your specific location... or you could just send a text message to Aardvak, and hope somebody answers.

Most enterprises probably would not need something like this... they would probably be better served with a dedicated "help desk" support staff to keep track of common questions and who knows what. However, I could see this being huge for the general public. It could potentially be the revenue stream that Facebook or Twitter have been waiting for. Or, it could be gobbled up by Google as a pre-emptive strike.

I'm just waiting for the bidding wars to begin...

Should You Trust Your Integrator To Pick Your Vendor?

CMS Watch was recently asked should I let my implementation consultants pick my ECM vendor? That's a pretty tough one to answer. On the one hand, your system integrator would have a fairly informed opinion about what software would work best for your problem. On the other hand, they may have a financial incentive to push you towards one vendor over another... so even if the vendor merely guides the process, there could be danger. A biased consultant will generate a biased RFP which may intentionally favor one vendor over another. I've seen it happen many times.

Hopefully I will not alarm anybody when I confess my bias is usually towards Oracle software. I have this bias for a few reasons:

  • they have good products that are easy to install and scale well,
  • I know their products the best, so I'll be able to implement an Oracle-based solution faster than a non-Oracle solution,
  • I have a great working relationship with Oracle

This last point is more important than most people realize... The CMS Watch article focused mainly on the negative aspects of a close relationship... but there are several positive ones as well. The main one being that enterprise software is typically very complex, with all kinds of hidden features and tricky undocumented configuration. If you or your implementation team has a close relationship with the vendor, you'll be able to extract much more value out of the software you purchased.

So, what's the "fair" solution? Kick out anybody with bias? Not quite...

I usually recommend that clients get unbiased feedback to do the initial selection... The biased person should not be involved here, because they may favor products that are a bad fit, simply because it's "what they know best." However, after the initial selection, the problems changes a bit. At this point, any of the top 3 should suffice... so the question now is which one will make your team the most successful? At this point, biases are ok, as long as they exist because your TECHNICAL team knows and loves the product. If they are familiar with it, comfortable with it, and know how to maintain it, then let them go with their gut... even if it seems like "bias."

You shouldn't be afraid of technology biases: they are there for a good reason... you just need to know when to keep them in check to avoid making a costly mistake.