Do we want to keep support for irregular splitfiles? An irregular splitfile appends several different keys (note that they cannot be USKs, at least, not without breaking the useful property that anything that isn't a USK is immutable) to get a final result. The sub-keys can be splitfiles themselves, but we can only append the entirety of the subkey; we cannot choose a small slice of it.
IMHO, if an API is provided for making these, they could be quite useful. You could for example create a tree-like structure on an HTML page, or allow the user to turn on and off each section one at a time, all while using very little space. However the code and the metadata spec would be simpler if we eliminate support altogether. It would be good to make the decision in the near future because I will be making some optimisations to the Metadata class soon. Obviously some XML processing format which can grab sub-sections of XML files, stick them together, and then transform them completely, would be better. However, that would be rather more work and won't be supported for a long time unless a volunteer comes forth (note that java has XML parsing support and even an XML transformations engine; it wouldn't be as hard as it sounds to have safe fproxy-side transformation support etc, but it's not critical path and won't be until at least after 0.8). So, does anyone think irregular splitfiles are worth keeping? -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20070216/c2828602/attachment.pgp>
