When I first started at ReadWriteWeb, the updates to Facebook were automatically posted via a Facebook application. It was an easy way to make sure our fans got to see our posts, but it didn't foster community discussions so after I got my bearings around here, I stopped the app (or at least I thought I did).

I'm a glass-is-half-full kind of gal, so rather than stress about the accident, I decided to look at this as an opportunity to compare apples-to-apples and determine which was better, auto-posted or manually posted items on Facebook.

As you'd probably guess, I found that automated posts saw significantly less views on Facebook. An auto-posted story that received x views on Facebook would receive, on average, 2.5x views on Facebook when published manually. Because more people saw the manually posted stories, their engagement was roughly doubled (likes and comments).

This does compare apples-to-apples because these were the exact same stories. The auto-posted content was posted first, within minutes of posting to our blog. To make sure that the data was representative, I went back a month and looked at average views, likes and comments and the auto-posted content's stats during this period is almost exactly in line with previous months' performance.

My assumptive explanation for this behavior is EdgeRank. EdgeRank is the algorithm that Facebook uses to determine where a post shows up in a user's stream. It seems to push content that gets reactions higher up the stream. I'm not sure if my manually posted entries got a better reaction because I worked hard to craft a pithy call-to-action on them (thereby moving them up in EdgeRank) or if some other secret sauce caused them to move up in EdgeRank, which in turn gave them more reactions.

Whatever the cause, the manually posted entries saw more than double the views and a little more than double the engagement. We did eventually get the app completely turned off and I posted the remainder of the month's stories to Facebook manually. This again showed a definite increase in traffic back to our site from Facebook (nearly double).

Manual posting is a chore. What takes the app seconds to post may take me 10 minutes. And, because I am not continually at the computer, some of our content isn't posted immediately after posting. There are definitely cons to manual posting, but the increase in engagement and page views back to our site is worth the additional labor.

Have you tested your Facebook posting methods? I'd be interested in learning if this obvious increase in performance was the same across other sectors.

This 99 cent app can take copied text or URLs (like links I've favorited on Twitter, for example) and apply an algorithm that shortens bodies of text thousands of characters long down to one thousand, five hundred or 140 characters automatically. How well does it work? It works well enough for me to appreciate it. Check out the demo video below.

"The algorithm finds the sentences in the passage of text that are most integral to the passage's meanings," says Trimit's Nick D'Aloisio. "We're utilizing keywords that are specific to the text, the appearance of adjectives and modifiers, where sentences are in the passage, use of conjunctions or contradictions, superlatives and dates, proper nouns, facts and figures, place and time adverbials, the syntax of the sentence (e.g if has list format) etc."

The end result seems good to me most of the time. I got the gist of Danny Sullivan's long post about the new Google Analytics interface delivered to me in 25% the time it would have taken to read the whole article, thanks to this app. ReadWriteWeb's Klint Finley thought Trimit did a poor job summarizing one of his articles but I thought the summary was just fine. I've been going through my Twitter Favorites, grabbing URLs, popping them into Trimit and loving it.

It's not as good as some people could do at summarizing the text, but it's much better than most people (humans) could do. And most of the people who could do a better job than the algorithm Trimit has built aren't available to me at a moment's notice.

This kind of technology could of course be applied in lots of different places and there are no doubt many different companies building things like this. Making it relatively easy to use on my phone and selling it for 99 cents is very nice, though.

D'Aloisio says the company will release Mac and iPad apps in the next two weeks, a web interface and bookmarklet in a few months.

Imagine something like this being built into Instapaper and fed automatically: please store offline everything I favorite on Twitter, all the most retweeted articles among my friends or on some other site, then create 500 word summaries of the articles and let me choose which ones I want to drill down into the full text of.

That sounds great, but Trimit is pretty cool already just as it is.

Let's test this baby out. Trimit says the above article is 2922 characters long. Cut down to just over 500, Trimit's summary reads as follows. What do you think?

Short form, long form, there's a time and a place for reading all kinds of articles but wouldn't it be nice if you could have some very long things made very short, automatically, and still get the gist of them. And most of the people who could do a better job than the algorithm Trimit has built aren't available to me at a moment's notice. It's not as good as some people could do at summarizing the text, but it's much better than most people (humans) could do. That sounds great, but Trimit is pretty cool already just as it is.

That looks like a succinct but accurate summary of the above review to me.

The Influences Driving Virtualization is the latest white paper from ReadWriteWeb. In the white paper, writer Philip Morgan looks at the influences that are shaping virtualization with particular emphasis on standardization, simplification and automation. Through that lens the white paper examines the factors that influence companies and how the cloud plays a defining role.

• Examine your risk to benefit ratio.
• Consider security implications.
• Learn best practices through enterprise case studies.
• Reduce your data center footprint.
• Outline standardized administration for better disaster recovery.
• Recommend strategic reallocation of resources.

We think you'll find our report about the influences driving virtualization will be essential reading for the new enterprise.

Using a system developed by Ariane Systems, guests check in prior to arrival using the "Allegro web/mobile check-in platform." They click a link received via email or text and establish check-in time, manage room preferences, update their profile and pay.

Thomas Westergaard, senior VP of Comfort Hotels for Nordic Choice Hotels said this was, in part, a matter of playing catch-up.

"The airline industry automated the check-in process 10 years ago, and we feel it is time that the hotel industry follows suit."

One of the make-or-break points on this experiment will be, as always, reliability. There is even less convenience in a time-saving measure malfunctioning than in its lack.

While most will agree that some automation is a boon, the disagreement may lie in what aspects of our interactions should be meat-free. Personally, I love not having to stand in line waiting for one distracted or chatty cash-register specialist for a pack of batteries, I have not found the automated airport experience anything less than panic-inducing.

How much of our travel experience do we want to appear convenient in order to save that last dollar, dinar or dalasi? These days, maybe every last vestige.

Circumstance did not allow for an ice hotel reference. I apologize.

MakerBot says the Thing-O-Matic prints higher-quality items than its other 3D printers but the real differentiator here is the automation: give this thing enough plastic to chew on and it will print the same or different objects one after the other after the other, clearing itself out each time before beginning anew. Can the Thing-O-Matic self-replicate - be used to make more Thing-O-Matics? That line was crossed by the old school MakerBot last June, so perhaps it's just a matter of time.

If you find yourself in New York City, you can presumably see all these things in the flesh at the new MakerBot Botcave retail store that opened on Black Friday.

Now if you'll excuse me, I need to go ask my Replicator to make me a cup of tea. Or not.

Where Are the Other Cloud Agents?

Back in December, we wrote about the rise of cloud agents (a term coined by Chris Arkenberg). These agents are automated applications that help us parse through the data swarming around us to provide us with the information we need. At the time, we highlighted a service similar to Tweecious called Twitchboard, another app that also posts the links you tweet to Delicious. According to the Twitchboard site, more services beyond Delicious were "coming soon," but here it is April now, and no others have been integrated. That's disappointing to say the least, but what's even more disappointing is that these sorts of "cloud agents" are so few and far between.

At the time of the previous post, some people missed the overall point, thinking we were raving about a Twitter to Delicious cross-posting app. The truth is, we were excited about this concept of automated agents. Unfortunately, since that time, we've seen very few apps that fit the description. In fact, only a couple of others really come to mind...and to be frank, we're not sure if they even count.

One such service is SocialToo, a "social web companion" which has the ability to automate many tedious Twitter tasks like catching you up on your follows, unfollowing people who didn't follow you back, deleting all your friends, or automatically following people for you. It's the service's automated, "set it and forget it" behavior that we think might allow it to be classified as a cloud agent. Another is "Twollo," a Twitter service that finds like-minded tweeters and follows them for you automatically.

Within the Twitter ecosystem, there are probably slews of others which we just can't think of right now. But do apps that automate tasks within Twitter count as cloud agents? They're not all parsing information to separate signal from noise (well, maybe Twollo is), they're just saving us time. We would love to see some "real" cloud agents emerge that provide more than just automation. We want intelligent cloud agents, too. Know of any?

What Twitchboard does is tie together different services on the social web and automates their interactions. Specifically, Twitchboard watches your Twitter stream and notices when you post a URL. It then automatically sends that link to your del.icio.us account. And, according to the company's homepage, they're working on connections to many other services for the future.

Blogger Chris Arkenberg says Twitchboard is a part of the "emerging class of cloud agents." These cloud agents, as he describes them, will help us sort and search the massive volumes of data we interact with regularly. He envisions that soon we'll have many of these cloud agents, swarming around us, working on our behalf, helping to parse the data flowing in and providing us with the information that we need, separated from the noise.

If that's true, then we seriously can't wait. We hope that in 2009, we'll start seeing more of these smart cloud agents and less of those "yet another"apps. Developers, start your engines, it's time to build some real tools.