Use droid signature files

[akkie]

Would it make sense to use the free DROID signature files to detect the mime type? They are completely free and actually maintained: https://www.nationalarchives.gov.uk/aboutapps/pronom/droid-signature-files.htm

[TonyValenti]

Hi @akkie - This is a great find! Do you understand the format of the XML files? If you can write some code to convert the XML to a Definition, I will certainly incorporate that!

[akkie]

Sorry, I don't understand the format. I searched for a mime detection lib and someone in the dotnet repo suggested to start with this source for a mime type utility. Then I stumbled over your lib and thought this would helpful.

[tiesont]

I'd imagine LINQ-to-XML can handle deserializing those definitions fairly easily. If someone still wants to explore this route, I could probably spend some time building a parser/loader, although I'm not sure how quickly I would get to it.

[TonyValenti]

I could bang it out really quick but I don't know what the different xml nodes mean. Do you have a documentation link?

[tiesont]

I was planning on starting with just being able to deserialize to a C# object, then figure out how those map to the relevant properties.

I was going to start here

https://www.nationalarchives.gov.uk/aboutapps/pronom/puid.htm

which does link to this:

https://www.nationalarchives.gov.uk/aboutapps/pronom/pdf/pronom_unique_identifier_scheme.pdf

That PDF at least describes the XML format, although I haven't dug into any of this enough to be able to answer any question beyond "is there anything out there explaining this format?".

[TonyValenti]

Hmm. Maybe I'm not understanding this but the PDF doesn't seem to describe the xml at all. What I see is a brief description on how the "ids" are basically organized.

[tiesont]

Like I said, I haven't dug too much into this. I know extracting the XML files into a usable format shouldn't be too difficult, and I think I see how to translate the nodes into the magic bytes and whatnot. I just can't explain any of it, at the moment.

Basically, I was volunteering to work on this, if someone still though it was useful.

[TonyValenti]

Yes. I definitely think this is useful.

Take a look at the existing code in the repo and how I deserialize the TRID signatures from xml. I would appreciate it if you followed a similar convention / process.

[TonyValenti]

This resource may help:

https://www.nationalarchives.gov.uk/aboutapps/fileformat/pdf/pronom_4_info_model.pdf

[TonyValenti]

Hi @tiesont . I just wanted to touch base and let you know that if you write the code, I will likely merge the PR.

[tiesont]

Hi @tiesont . I just wanted to touch base and let you know that if you write the code, I will likely merge the PR.

Sorry, other stuff has been keeping me busy. Still happy to take a crack at it, but it might be a little while.

[ArcticLampyrid]

This may explain how to use droid:
https://github.com/digital-preservation/droid/wiki/files/DPTP-01.pdf