archivematica / Issues

Issues repository for the Archivematica project
GNU Affero General Public License v3.0
16 stars 1 forks source link

Problem: The FPR uses 'enabled' to mean two different things for format identification commands #860

Open evelynPM opened 5 years ago

evelynPM commented 5 years ago

Expected behaviour In the Preservation Planning tab, information about whether or not a tool is enabled should be consistent

Current behaviour See attached screenshot. Both "Enabled Yes" and "Enabled No" show for the same tool, and under Actions there is only the option to Disable, whether or not the tool is currently enabled.

Capture

Steps to reproduce In the Preservation Planning tab, go to Identification > Tools > and click View for any of the tools.

Your environment (version of Archivematica, OS version, etc)


For Artefactual use: Please make sure these steps are taken before moving this issue from Review to Done:

sromkey commented 5 years ago

Now that we're managing the file ID commands in the FPR instead of in the processing configuration, I wonder if it even makes sense to display the tools at all? It would be more consistent with other preservation actions to just remove it (the commands can be enabled/disabled instead).

evelynPM commented 5 years ago

That sounds like a good idea, so long as it's not too cumbersome to switch format ID tools. I recently dealt with a use case where fido was better than siegfried at identifying jpeg images with a particular colour space. So the user wants to be able to switch to fido when those types of jpeg images are being ingested, but use siegfried for other filetypes.

ross-spencer commented 5 years ago

@evelynPM that sounds suspiciously like a bug in Fido. I know there are a few quirks in both that we can use to our advantage, e.g. the media samples we use for something SF can't handle. But both should be giving the same result as PRONOM as a standard. If you happen to have samples i can investigate and log that with one of the tools or PRONOM.

evelynPM commented 5 years ago

It would be a bug in siegfried, actually. It was fido that did the identification correctly. I can share some files with you.

ross-spencer commented 5 years ago

That'd be great @evelynPM thank you! yep, I shouldn't jump to conclusions too quickly.

sromkey commented 5 years ago

Did I misunderstand the problem... I thought it was just about how it's enabled/not and displays in the FPR page? Edit- @ross-spencer I realize now you were probably referring to the comment above about reasons for enabling/disabling one tool or another.

sromkey commented 5 years ago

Regarding enabling/disabling the commands, it's not cumbersome at all- if you enable the one you want, the others are automatically disabled.

sallain commented 5 years ago

Just a note on this - I think that the two "enabled" fields that the screenshot above shows mean different things. All three tools are always enabled (i.e. they're always available to Archivematica), but only one command can be enabled at a time. The top "enabled" field refers to the tool, the bottom "enabled" field refers to the command.

Screen Shot 2019-08-19 at 4 15 57 PM

In this example I have File Extension enabled, so it shows that both File-Extension-the-tool and the command to run it are enabled:

Screen Shot 2019-08-19 at 4 15 44 PM

I have Siegfried disabled, so while Siegfried-the-tool is enabled, the command to run it is disabled:

screenshot2

These two states remain accurate if I change up which command is enabled on File Identification > Commands.

So, this isn't a bug, which is good, but it is definitely confusing for the user.

A short-term solution could be to just remove the command section from the tools page - it's not necessary, and doesn't exist for any other tool in the FPR. That said, I agree with @sromkey that the identification tools shouldn't be managed any differently from the rest of the tools in the FPR, so probably the whole File identification > Tools page should go, with the FI tools managed via the Format policy registry > Tools page instead.