Yeah, so Bespin a cool editor, and it makes use of lots of bleeding edge tech, (and it's got several annoying bugs - sorry about those). But I think Bespin has potential way beyond just looking nice ...

Open source development: You hear about project X, use it, and it’s nice, but to use a worn out phrase - there’s an itch to scratch ...

So what do you do?

Open Source Development Before Bespin

You search the project website for the directions to the source control system, you check that you’re up to date with cvs/svn/git/hg/whatever and checkout the source. Next you need to figure out how to build the project with ant/make/maven/paver/rake/whatever, and maybe download new libraries to make the compile clean and you then configure your IDE or editor to the new project. You can then go about editing the code. When you’re done with that you might need to figure out xunit/yunix/zunit to run the unit tests. Then you have to join a mailing list to work out how to send your changes in, which might be a patch or a bundle or a zip or direct commit access. You might discover that you need to sign a CLA so the project knows who you are, so you mess about with printing/signing/scanning/emailing/whatever.

And then you’re done. With complex projects there could be lost more to it.

While you’re doing this you have no idea what the other developers are doing. Maybe they’re working on a totally different branch. Maybe they’ve done something similar already. Maybe they’re making changes that mean you should try a different approach. It would be good to know because you might be wasting your time.

Open Source Development After Bespin

You visit Bespin, and elect to edit project X. When you’re done, you click submit, and the project owners get your change.

Plus you can see what the project owners are doing with the source, and how your changes fit in.

The Power of Bespin

For me, what I find interesting about Bespin isn't the tech, it's the lowering of the barriers to entry and the socializing of open development.

Virtually no software is perfect, and there are billions of people that are affected by the imperfections. There are hundreds of ways to contribute to any development project, and millions of people who could use their skills to help if only we could find ways to get the sources of help to the problems needing work

This application of huge resource to complex problems is what the Internet has always been good at, but so far we've not really done much to help software development. I hope that Bespin can become part of the solution.

This isn’t going to become a Bespin blog (that’s here or we're on twitter). Irregular DWR/Ajax/Security programming will return to my blog and twitter feed shortly.

Responses[8]

Posted by Joe Walker on 16 February 2009 12:38:54 GMT #

I’m very excited to be going to work for Mozilla Labs.

I have very much enjoyed my time working at SitePen - they are doing awesome work with Dojo and DWR, but with the main development on DWR 3 being done, it's time to move on. SitePen has some amazing talent, and it's been a huge privilege for me to have been able to share the ride, so I'm very grateful to Dylan and everyone at SitePen for allowing me to be part of it. I'm not going to be ignoring DWR either, so between me and the excellent team of Mike, Jose, David, Lance and many others I'm sure it has a long future.

My new role will be working with Ben, Dion and Kevin on Bespin.

You might be able to get hints about Bespin from postings that are around already. Oh, and the name - that might be a clue too. Needless to say I’m excited about the project. More soon.

Responses[14]

Posted by Joe Walker on 04 February 2009 11:27:33 GMT #

Recently:

• Aaron Boodman - "I'm a big believer in silent autoupdates"
• PPK/Quirksmode - "I hate automatic updating"

Something Aaron didn’t mention - What happens when a silent auto-update breaks something? Anyone that's had to do battle with the demo-pixies silently changing something just before an important demo hates the idea that something might be silently altering their rendering engine as they are opening their laptop on-stage.

PPK's issue basically comes down to testing older versions - how can I test version X-1, when it's just been replaced by version X. On the one hand with silent auto-update virtually everyone will be on version X, so X-1 isn’t important, however I don't think that removes all need for testing on X-1.

Solution?

We should have silent auto-update. It makes the Internet safer, and we badly need that. But we should also have a system that keeps old versions around for some period of time with a way for users to revert. Maybe there is then an entry in the help menu that allows users to say 'it broke take me back', and maybe a preference that configures the number of old versions left lying around.

I have a suspicion that many corporates turn auto-update off in order to test the updates batches before internal release (or to put it another way - it's a cost saving) Maybe with undoable silent autoupdate they would allow the system to work without getting in the way?

I'm a big believer in undoable silent autoupdates.

Responses[9]

Posted by Joe Walker on 04 February 2009 09:49:23 GMT #