LDraw.org Discussion Forums

Full Version: LDraw OMR website
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12
Oh, there's no hurry.
I just meant that there is a discussion about the (working of) site itself, not really the hosting Wink My phrasing was a bit weird.
yes, and to me the requirements for an OMR-compliant model are by far too restrictive to make it fun for me to contribute, sadly.
for me, it would be sufficient, if the file contained the model properly built, and its filename and title make somewhat sense.
this minimum set of req's would be enough for me.
of course, anyone wanting to go further formally inside the model, of course still may, but for me, the effort would be too high because I cannot see the gain.
I fully agree.
May I suggest a new idea?
There already exist several uploaded models all over the world.
My idea is that our OMR repository is not a model repository, but instead simply a database pointing to models stored
wherever else. This _could_ include a server by us own, but does not have to.
The DB could point to models uploaded wherever, be it brickshelf, eurobricks, ldraw.org, or wherever else.
Inside the DB, each model is represented by a URL where it is physically stored, plus a md5 checksum of that file.
For every such model, attributes can be stored like
- the model is built properly (and this has been verified by X persons)
- all stickers or patterns are properly present
- a list of known issues like missing or substituted parts
- from which set that model was built
- year when that set appeared from TLG
and so on

The frontend of our database then would allow to search for models like this:
"give me all train sets between 1980 and 1985 containing at least 20 parts which got reviewed by at least 2 people to be correct"

This suggestion is a fundamentally different approach than creating a "PT" for models.
It allows better integration of all existing models, without having them to be copied to our servers.
This avoids the problem to ask all the existing model authors for their permission.
At the same time it avoids tedious re-creation of non omr-compliant models,
and has the charm that plenty of models are available from the start (most of them un-reviewed yet, of course)

play well

PS, I agree with that we should drop the term "official".
I suggest titling that new thing a LEGO model-searching database.
If you want, it will in the end be a "peeron2", but with correctly built models, not only part lists.
The database could additionally match the uploaded models against existing part lists.
In the end, the new DB will be a much more powerful DB than peeron was, and this way also solve the problem of the slowly dying peeron.
Ok. The webspace is setup and a (maybe temporarily named) subdomain of omr.ldraw.org is created. From here you'll have to send me the source code via email (orion at ldraw dot org) or upload it yourself via sftp (if you know how to do this). Either way you'll have to email me for login credentials.
The only thing I consider as very important in the OMR specification is the latest added bit: inclusion of unofficial parts!
Well, what I have made so far is exactly what you describe except for the fact that the files are locally stored and not linked.

I do like the idea of linked files to create a bigger collection, but we have no control at all about the files hosted at other locations. A lot of files are hosted at Brickshelf, how long does Brickshelf still exist? A lot of files are hosted at Dropbox, what if a user deletes or moves the files in dropbox? etc. etc.

So, I highly doubt if it would be useful to include links instead of adding the files themselves. It could become very annoying when you'd like to download a file a year from now and the link doesn't work anymore.
Totally agree: while links to external files might be interesting, I think that the core of the OMR should be locally stored.
Although this thread had some discussion of "tagging" it related more to official information, such as year and theme.

It might be nice to also include freeform text tagging, so users can search for useful common terms like "building", "vehicle", "aircraft", "4-wide", and so on. Wouldn't want the tagging to get too out of hand (like the IMDB tags!) but perhaps up to 50 or so tags that are commonly used throughout the LEGO community.

It may seem like overkill now, but once there are 10000 models in there... Wink
Well, it's a nice idea for searching lego-sets, but like I said before we don't really need to reinvent the wheel. When looking for a set, you can better use Brickset (which has those tags) or Rebrickable and once you found a set for which you'd like an LDraw file, you can go the OMR website to download that model. It would be a bit waste of time to make yet another database of Lego sets. That's my view of it at least.

Of course both Brickset and Rebrickable have an API for set info and maybe we can do something with that in the future, but let's get the thing online first Wink Any progress Orion?
It seems we're life! Smile
Thanks Orion!

At the moment, there's still a little problem with the (albeit small and simple) admin interface, but the site functions fine for 'normal' users.

There are only 6 files at the moment. I still can't download the zip file with current OMR files, the ldraw.org site still throws an error. Does anyone have an additional download link?

Let me know what you think about it (either good or bad). Also, if you think you can write something better for the homepage/about page, feel free. Writing text is certainly not one of my best skills...

Btw, I thought I changed all OMR to MR (Model Repository), but the huge heading at the front-page still says OMR. Oops...
Pages: 1 2 3 4 5 6 7 8 9 10 11 12