we need a "comments" section at the OMR

[Steffen]

I spotted various errors in various OMR files at omr.ldraw.org.

Currently there is no way to report or correct them.

I would love if we could add (even a simple mechanism would suffice) something which allows us to correct OMR errors.

This could be
- a comments section below a displayed set which allows to add remarks
- an upload button which allows to upload a corrected version of the file
- a "parts tracker"-like mechanism which tracks OMR models and allows reviewing them
- a button to send a message to the author
- or whatever else.

Currently, I feel forced to be very passive over there, seeing all the problems, but no way to correct them.

What can be done? What do others think?

[Orion Pobursky]

I spotted various errors in various OMR files at omr.ldraw.org.

Currently there is no way to report or correct them.

I would love if we could add (even a simple mechanism would suffice) something which allows us to correct OMR errors.

This could be
- a comments section below a displayed set which allows to add remarks
- an upload button which allows to upload a corrected version of the file
- a "parts tracker"-like mechanism which tracks OMR models and allows reviewing them
- a button to send a message to the author
- or whatever else.

Currently, I feel forced to be very passive over there, seeing all the problems, but no way to correct them.

What can be done? What do others think?

Right now, everything on the OMR is added manually from the forums. I think it should stay that way. The OMR is intended to be the "library" and the PT equivilent to the OMR is here on the forums. In other words, the models should be posted here on the forums first and that is where I feel discussion should happen.

[Orion Pobursky]

We could, however, add a link back to the relevent forums post.

[Merlijn Wissink]

Well, a long time ago (in a galaxy far far away) I considered the idea of adding user-upload so the OMR would be less depending on the forum. I scrapped the idea however, since then there would be no control at all over the file (quality). And if I remember correctly most people agreed. But, to my own surprise, I never thought about something part-tracker like where models have to be reviewed.

Implementing something like that would be pretty nice, but is quite a lot of work though. I'll look into it, but I can't promise anything. A simple 'report problem' button is probably a lot simpler.

[Steffen]

May I suggest a radical but simple idea here:


Idea: use the existing PT for OMR models.

they would be submitted into a new folder "omr", or, what I would like even much better, into "models".
(because that's what they are.)

We then could use the normal existing submit + review mechanisms of the existing PT for checking them.

When the parts tracker is used to make a release, the "models" folder simply will not be included in the final release.
Instead, a separate "omr" (or better: "models") release could be added in a later step.

[Steffen]

>  I scrapped the idea however, since then there would be no control at all over the file (quality).

Right now you have the opposite effect:
- there are mistakes in the OMR files which don't get reported
- the submittal process is manually tedious (collecting from forum posts, tracking where the corrected versions are, etc.)

So you now have minor quality _plus_ tedious manual work.

Having a self-submitting and review mechanism, you could drop the handwork, _plus_ would get better quality.

I think the absolute minimum would be a "comments" section at each model, where spotted problems can be reported,
and marked with a checkmark if the model author has resolved that problem.

[Steffen]

We need version control for the models on the OMR:
there must be a mechanism to replace a flawy model by a new version.

The ideal mechanism to achieve that would be a git repository.
It even could be stored on github.
Then people could clone it from there and get all the models including their history.

The OMR site at ldraw.org could simply be a graphical frontend for that repository.

[Merlijn Wissink]

>  I scrapped the idea however, since then there would be no control at all over the file (quality).

Right now you have the opposite effect:
- there are mistakes in the OMR files which don't get reported
- the submittal process is manually tedious (collecting from forum posts, tracking where the corrected versions are, etc.)

So you now have minor quality _plus_ tedious manual work.

Having a self-submitting and review mechanism, you could drop the handwork, _plus_ would get better quality.

I think the absolute minimum would be a "comments" section at each model, where spotted problems can be reported,
and marked with a checkmark if the model author has resolved that problem.

Yes, I certainly agree with you there on all points. The thing is that when I created the OMR site it was just a small side-project to gather some experience and I wanted to keep it simple. The fact that it actually became an LDraw site was never really planned at the beginning. Since then I learned a lot more and I'm confident that I/we can create everything we would like to see in an 'OMR 2.0'; the only rare good is time. The more we want, the more time it takes...

But, the ideas you bring to the table such as using the part tracker as review platform might take less time. I have no idea however, if something like that is possible from a part-tracker standpoint.

[Orion Pobursky]

May I suggest a radical but simple idea here:


Idea: use the existing PT for OMR models.

they would be submitted into a new folder "omr", or, what I would like even much better, into "models".
(because that's what they are.)

We then could use the normal existing submit + review mechanisms of the existing PT for checking them.

When the parts tracker is used to make a release, the "models" folder simply will not be included in the final release.
Instead, a separate "omr" (or better: "models") release could be added in a later step.

I am stridently against any type of PT like review process. I’d rather have flawed models in the OMR than have them go through the review hell that Parts are subject to. For the OMR I’d rather err on inclusion as opposed to the PT which errs on exclusion. That’s said, I’m not opposed to some sort of objective grading system but has to be automated and not subject to interpretation.

[Philippe Hurbain]

Fully agree with Orion...

[Willy Tschager]

I am stridently against any type of PT like review process. I’d rather have flawed models in the OMR than have them go through the review hell that Parts are subject to. For the OMR I’d rather err on inclusion as opposed to the PT which errs on exclusion. That’s said, I’m not opposed to some sort of objective grading system but has to be automated and not subject to interpretation.

+1

w.

[Michael Horvath]

We need version control for the models on the OMR:
there must be a mechanism to replace a flawy model by a new version.

The ideal mechanism to achieve that would be a git repository.
It even could be stored on github.
Then people could clone it from there and get all the models including their history.

The OMR site at ldraw.org could simply be a graphical frontend for that repository.

Does Github allow you to comment on each individual file? Otherwise it is no better than the forum.

[Steffen]

No, I meant to suggest github as a possible backend for a future "OMR tracker".
It would be used to store and update the OMR model files.
And that would happen by some web frontend which works like the PT.
This way the effort needed would "only" (...) be to write the web frontend implementation,
but not to store all the model files.