OSS Watch team blog

Replay is an open solution developed in java to integrate, index, archive, and distribute audiovisual lecture recordings on a large scale. It’s being considered as the initial code to support the OpenCast community. The Opencast community is a collaboration of higher education institutions working together to explore, define, and document podcasting best practices and technologies.

As I’m currently at ApacheCon I am in geek mode, so I figured I’d have a play with the 0.4.1 release of Replay. I tried to install it before I left for the airport but the installer failed. I had time to check out the Replay issue tracker and found that this was a known issue and there was a hint in there as to how to solve the problem. However, this required me to download a 1Gb file from Apple. Since my taxi was due to arrive I had to give up on the idea of getting it installed. So I updated the issue with my findings and started my journey.

Over the last few days I’ve been working on this in the odd five minute here and there - how many conferences do you have sufficient bandwidth to download a 1Gb file in a reasonable amount of time?

I discovered that there were two other issues that would prevent things working - both appeared in the FAQ, but I can be pretty sure that both of these issues will apply to the majority of Mac users (as opposed to Mac developers), so I hope these make it into the install documentation. Furthermore one of the issues, relating to proxies, was not correct (if you care I had to remove both entries in the proxy settings, not just the one identified in the FAQ).

I was now able to upload new content to the system - cool.

The idea of Replay is that the system does speech and text recognition to make the video searchable. However, it didn’t look like this was working, searches didn’t produce any results. So I decided to go for lunch, a shrimp and oyster Po Boy from the excellent Acme Oyster House (I’ve eaten so much shrimp whilst here in New Orleans - this town knows how to cook shrimp!)

When I returned from lunch I sat down to have another go and discovered that  the search page was now showing a tag cloud. It seems the indexing had worked, it’s just that the UI didn’t tell me it was in progress earlier on.

Very early impressions of the system are that it is very promising - I love the fact that my podcasts can now bew indexed automatically. But I haven’t actually played with Replay enough to see if it works that way.

Although I had some problems getting it running I was, with a little work, able to find the information I needed. It’s perfectly acceptable for a 0.4 release to have teething issues like these, hopefully this feedback will help the team iron them out. It’s always hard for developers to know which issues users will hit and feedback like this is vital (note without the feedback of previous users I would not have been able to get this install working).

I’m going to be playing with adding content over the next few days and will report further in the future.