Peder Borg Borup

Currently working a lot on the podcast worker. As I’ve mentioned before, the worker will handle several things:

• Cutting up a medie-file into smaller items
• Adding the to a database
• FTP’ing the media-files to an Anystream Transcoder
• Analyzing the feedback from the Anystream Transcoder and make the items transcoded publicly available

Apart from that, I also need to create the websíte, where people chooses which categories (e.g. sports, news, regional) they wan’t to subscribe to. And of course create the engine that generates the RSS on demand.

Until now, I’m almost finished with the internal workflow (the parts that are invisible for the end user).

I’ve:

• Used the FileSystemWatcher to guard, that files are not transfered to the Anystream Transcoder while they are still being written by the process that cuts the main file into smaller files
• Created a data abstraction layer using stronly typed datasets (really like this) againt a MS-SQL-database
• Created a FTP-program that contacts a webservice, which starts up the whole process

I’ve calculated, that 8 serveres are a part of this entire podcast-process (from mediafile/news-item to webserver). I do know, that a lot of you guys do very large projects – but to me, this is pretty heavy.

And most of all; I have to be extremely sure, that all errors are handled, as the process has to work perfectly without people “helping”.

I look forward to the day it works (which is about 11 days when we go into beta).