On a side note, in a recent Seesmic Desktop update, the application not only caused my plugin not to work, but the entire application crashes altogether. I’ve made the appropriate changes to my plugin to prevent this from happening but just to comment on it… it was a change they made, not me that caused this and I’m still discussing with them what could have caused it. So if that happens, you need to download the new XAP file here and manually put it in your Documents\Seesmic\Plugins folder.

So, in my case, if you installed selecting only 2 of 4 features and then did an upgrade on that install, you’d get all 4 features installed after upgrade. Once I modified the condition of this action to not set the InstallLevel during upgrades, post-upgrade, you only get the 2 features you originally installed with.

Just wanted to get this info out there since it wasn’t easy to figure out…

Go here to download it and let me know how it goes!

Let’s start with what’s out there that’s the most popular mentioned OCR library available, Tesseract. I see on the Google project site that this was one of the top 3 OCR engines in 1995. I feel sorry for people in 1995. I’ve got experience with commercial OCR engines like LeadTools and even the simple one like Microsoft Office Document Imaging (MODI) and they’re much better than Tesseract.

Then there’s SimpleOCR. They’re closed source but freeware. Thanks to those guys for providing a freeware and royalty-free SDK but man, it’s absolutely terrible. In their defense, I don’t think it is maintained anymore (it doesn’t even support color) but could take a screenshot of the letter S and this library would tell me it’s a T. Bleh. I’ve heard MODI is powered by the same engine as SimpleOCR but if that’s true, you’d expect similar results which you don’t get.

Now here’s one I want to try – Microsoft Research Project Hawaii, OCR in the Cloud. I don’t know how fully featured or how dependable this is since it’s a research project but as soon as I get some free time, I’m gonna try it out. I’ve tried WiseTREND’s OCR Cloud service and I love it… it’s fast and super accurate (uses their ABBYY engine) and also supports a lot of file types… but if I want to distribute an opensource desktop application that doesn’t a) require internet connectivity and b) more importantly doesn’t require per document pricing, WiseTREND is obviously not an option.

Just a rant on OCR products… maybe I’m wrong about some of things I said and maybe there are better alternatives out there…. which hopefully both are true and some commenters will lead me on a better path. Overall, OCR in the Cloud is great because of how expensive OCRing (and ICR even more so) is on any processing machine so it’d be great to see more options of this nature but my only problem from the litigation world is that I don’t think lawyers are exactly comfortable sending privileged client data over web for this so in a lot of cases, desktop-only solutions are the only solution.

Anywho, enough gushing over how awesome the guys behind Seesmic are… I’ve updated the plugin yet again this time adding migre.me short URL support. Thanks to bit.ly and migre.me for providing easy APIs to expand short URLS… and to is.gd, you need to get with the program and provide this service too!. I also hopefully enhanced performance a little by multi-threading the actual creation of the thumbnails themselves so there may be a half-second delay in their creation but you’ll thank me when you see 10 posts in a row with video links and the app doesn’t slow to a crawl trying to create the thumbnails.

You can get the newest version here.

As for the previously mentioned issue around auto-updating, I’m doing everything right and SD2 is properly checking the right URL but it’s still not working… I hope it’s just a caching thing on my local machine so hopefully just an issue with me and not everyone else. I’ll update on this situation if I see any change…

EDIT: It’s working now! Forgot to put the build number for the Platform version.