3 ms·

Building an Internal Social Network Your Company Loves

3 points by mankins 14y ago

mindcrime 14y ago
Excellent article, a lot more "meat" than I expected, given the title and the usual level of discourse around this topic. That said, I'd advocate considering starting an internal social network project using an Open Source foundation. And, if you'll indulge me a moment of self-promotion, I'll point out Fogbeam Labs' suite of projects[1] as a good place to start - especially for a firm who wants to help shape the direction of the product(s) they will be using, but don't necessarily want to be responsible for maintaining an ESN product for perpetuity. The thing we believe sets our approach apart from a lot of the other ESN products out there (Yammer, Jive, etc.) is our approach to integrating business processes and workflow with the "social" aspect of the ESN. It's still very early, not everything is implemented, but some really awesome stuff is in the pipeline. And to be fair, there are some other OSS ESN projects out there as well. There's always Elgg, and Alfresco have been adding some "social" aspects to their Document Management product. [1]: http://www.fogbeam.org/ http://www.fogbeam.org/
- mankins 14y ago
  The openness sounds interesting. What kind of stuff is in the pipeline?
  - mindcrime 14y ago
    There's a ton of stuff coming, but some of what I'm most excited about include: 1. Using machine-learning (via Mahout) to find correlations between people, events, and content. We have some fairly naive content <-> content correlation stuff now, but Mahout lets you do a lot more and better stuff, and it's more scalable thanks to being built on Hadoop. 2. More and better support for the "Business Events Subscription" and "Workflow Integration" stuff. We already have support for tapping into an ESB and consuming XML messages and subscribing using XQuery expressions. But we currently only support one vocabulary (OAGIS), so we want to do several things with this: a. Support some other "stock" vocabularies b. Support arbitrary (customer specific) schemas / vocabularies c. Support non-XML messages d. Build a graphical subscription builder of some sort. Right now you have to know XQuery and manually write an XQuery expression to do a Business Event Subscription, and you obviously can't expect average end users to know XQuery. 3. We also plan integration with some workflow engines (Activiti and/or jBPM will probably come first) so you can see your User Tasks in your stream (if you want to, that is) 4. There's a big UI refresh coming very soon as well. 5. We have some stuff in the works regarding using Semantic Web technologies to start to really make this more part of a "knowledge management" approach than it currently is. We want to support making semantic assertions about people, events, things, content, etc., and allowing queries as rich as what you can do with SPARQL. But, the challenge is to take that kind of power, and simplify it so that you aren't asking everybody to learn SPARQL, which - less face it - isn't going to happen. Just like with XQuery, you want the power it brings, but without overwhelming the average users. That's one of the big challenges we have in front of us, actually... finding the best way to provide power without scaring the users off! :-) 6. Integrating Apache OpenMeetings to support collaborative video-chat, whiteboarding, etc. is on the table as well. We've started playing with this, but it'll be a while before it's ready, if we commit to doing it. 7. There's also a tentative plan to add support for OpenSocial widgets. Interestingly enough, OpenSocial - to the extent that it has any traction at all - seems to be more popular in the enterprise. Could be something to do with how Facebook (obviously) never adopted it, and then Google never implemented it on their own social network (the one that matters, not Orkut). There's a lot of other stuff on the roadmap as well, although some of it is in the "unplanned, we'll get to it one day" category at the moment. Since this is an enterprise play, we're looking at things like integrating with HRM systems, so that your "profile" on the ESN is automatically populated with the "stuff" that is in HRM (obviously this is only for companies that employ a dedicated Human Resources Management system). Something else we'll probably do, is build some adapters for generating ActivityStrea.ms messages from other systems that don't currently support emitting those. We can already consume ActivityStrea.ms, but that isn't as useful if other apps don't push those messages out. Some of this might be done on a "one off" basis if we get a customer who asks for it, but we might just pre-build such an adapter for something popular that we expect to encounter a lot of. Anyway, I could ramble about this for a while, and this has gotten longer than I meant for it to already. I think some of the stuff we have coming is going to be very cool and very useful, but we're pretty resource constrained at the moment, so progress is never as fast as we wish. That said, it's all Open Source and we welcome outside contributors. We've built on the backs of a lot of existing great work (Groovy, Grails, Lucene, Tika, Rome, Mahout, Camel, Shindig, PostgreSQL, etc.) and are committed to being a true Open Source project... not one of those "Open Source In License Only" deals where a company develops something behind a wall and occasionally throws a code dump over the wall.

3 points by mankins 14y ago

Excellent article, a lot more "meat" than I expected, given the title and the usual level of discourse around this topic. That said, I'd advocate considering starting an internal social network project using an Open Source foundation. And, if you'll indulge me a moment of self-promotion, I'll point out Fogbeam Labs' suite of projects[1] as a good place to start - especially for a firm who wants to help shape the direction of the product(s) they will be using, but don't necessarily want to be responsible for maintaining an ESN product for perpetuity. The thing we believe sets our approach apart from a lot of the other ESN products out there (Yammer, Jive, etc.) is our approach to integrating business processes and workflow with the "social" aspect of the ESN. It's still very early, not everything is implemented, but some really awesome stuff is in the pipeline. And to be fair, there are some other OSS ESN projects out there as well. There's always Elgg, and Alfresco have been adding some "social" aspects to their Document Management product. [1]: http://www.fogbeam.org/ http://www.fogbeam.org/

mankins 14y ago
The openness sounds interesting. What kind of stuff is in the pipeline?
- mindcrime 14y ago
  There's a ton of stuff coming, but some of what I'm most excited about include: 1. Using machine-learning (via Mahout) to find correlations between people, events, and content. We have some fairly naive content <-> content correlation stuff now, but Mahout lets you do a lot more and better stuff, and it's more scalable thanks to being built on Hadoop. 2. More and better support for the "Business Events Subscription" and "Workflow Integration" stuff. We already have support for tapping into an ESB and consuming XML messages and subscribing using XQuery expressions. But we currently only support one vocabulary (OAGIS), so we want to do several things with this: a. Support some other "stock" vocabularies b. Support arbitrary (customer specific) schemas / vocabularies c. Support non-XML messages d. Build a graphical subscription builder of some sort. Right now you have to know XQuery and manually write an XQuery expression to do a Business Event Subscription, and you obviously can't expect average end users to know XQuery. 3. We also plan integration with some workflow engines (Activiti and/or jBPM will probably come first) so you can see your User Tasks in your stream (if you want to, that is) 4. There's a big UI refresh coming very soon as well. 5. We have some stuff in the works regarding using Semantic Web technologies to start to really make this more part of a "knowledge management" approach than it currently is. We want to support making semantic assertions about people, events, things, content, etc., and allowing queries as rich as what you can do with SPARQL. But, the challenge is to take that kind of power, and simplify it so that you aren't asking everybody to learn SPARQL, which - less face it - isn't going to happen. Just like with XQuery, you want the power it brings, but without overwhelming the average users. That's one of the big challenges we have in front of us, actually... finding the best way to provide power without scaring the users off! :-) 6. Integrating Apache OpenMeetings to support collaborative video-chat, whiteboarding, etc. is on the table as well. We've started playing with this, but it'll be a while before it's ready, if we commit to doing it. 7. There's also a tentative plan to add support for OpenSocial widgets. Interestingly enough, OpenSocial - to the extent that it has any traction at all - seems to be more popular in the enterprise. Could be something to do with how Facebook (obviously) never adopted it, and then Google never implemented it on their own social network (the one that matters, not Orkut). There's a lot of other stuff on the roadmap as well, although some of it is in the "unplanned, we'll get to it one day" category at the moment. Since this is an enterprise play, we're looking at things like integrating with HRM systems, so that your "profile" on the ESN is automatically populated with the "stuff" that is in HRM (obviously this is only for companies that employ a dedicated Human Resources Management system). Something else we'll probably do, is build some adapters for generating ActivityStrea.ms messages from other systems that don't currently support emitting those. We can already consume ActivityStrea.ms, but that isn't as useful if other apps don't push those messages out. Some of this might be done on a "one off" basis if we get a customer who asks for it, but we might just pre-build such an adapter for something popular that we expect to encounter a lot of. Anyway, I could ramble about this for a while, and this has gotten longer than I meant for it to already. I think some of the stuff we have coming is going to be very cool and very useful, but we're pretty resource constrained at the moment, so progress is never as fast as we wish. That said, it's all Open Source and we welcome outside contributors. We've built on the backs of a lot of existing great work (Groovy, Grails, Lucene, Tika, Rome, Mahout, Camel, Shindig, PostgreSQL, etc.) and are committed to being a true Open Source project... not one of those "Open Source In License Only" deals where a company develops something behind a wall and occasionally throws a code dump over the wall.