In the post you write "A standard wiki isn't suitable for this project,
because of their 'everything goes' permissions model." I think that
you are wrong about the permissions model of the wiki. In most wikis there
is a very extensive mechanism to regulate permissions, starting from locking
single pages or groups with a keyword and ending in full username/password
authentication, including OpenID support. In fact my experience suggests
that wikis have even better permissions control than CMS systems. I do not
want to say that this makes wiki more suitable, as there are other requirements
to consider, but I think your argument against wiki is false, and that it makes
sense to look at wikis as a serious possibility. Below is a link to a PmWiki
page about user authentication, if you are interested to try it out fast I can
email you a slightly modified pmwiki installation that I use as a template for
new wikis; it defaults to requiring user authentication.
[www.pmwiki.org]
I find that once you have authenticated someone in a wiki (particularly by username, but even just by IP address) and you can tie their edits together, the effects of a malicious or clumsy user can be minimized. People claim that they don't want people editing their page, but they might soften their stance when they see someone fix a particularly egregious typo, or add a link to a pertinent presource. The most important thing is to keep every old version of a resource, and give users the ability to undo someone else's changes. Giving a transient reader the ability to add to potentially any page fosters more contribution than the "Do Not Edit Me" message.
Since the previous commenter made a wiki suggestion, I'll put in a plug for MediaWiki; it has many, MANY useful extensions available for it at mediawiki.org, including category- and article-based permissions and anti-spam extensions. In addition, many would-be contributers would already be familiar with the software that underlies Wikipedia. It also nicely handles file description and versioning.
I would recommend avoiding a hierarchical file archive. I think the better model would be one that allows tags or categories to be applied to files, rather than fixing one specific hierarchy. Users will be able to quickly generate and change as needed a good categorization system, and new users will generally adopt the categorization patterns they already see. It also facilitates flexible categorization - where do you put "houseboat", in "domiciles" or "nautical vehicles"?
RepRap is an exciting project, and a good object repository will keep people involved in the project. Good luck!
I just realized that you already use Mediawiki (that will teach me to not click on links before I start typing!) A quick anti-spam fix is to use the reCAPTCHA plugin for Mediawiki (http://recaptcha.net/plugins/mediawiki/), which provides captcha derived from scanning books; every edit will contribute to better OCR!
One advantage of staying with a wiki, especially mediawiki, but with stricter editing and authoring criteria, would be you could probably use the existing pages without conversion or alteration.
I'd say go with captcha of some form, along with an occasional review, by trusted editors, to kill those accounts that slip in anyway.
From the last days of editing, I noticed the early posts of spammers tended to be a very few lines of gibberish appended to the first line of a page. This, to me, suggests that all our spammer accounts were hand-generated, and then handed off to a spam program. Additionally, some of the posting behavior suggested several of the spamming accounts likely originated from the same point. I say this without access to the IP logs.
Here are two examples to illustrate.
[ite_support&oldid=7002" rel="nofollow" >objects.reprap.org]
This is an early post by a spammer. It looks like a test post to me, as 14 days later it was posting link-spam.
Both of the above pairs frequently posted within minutes of each other, suggesting to me they originated from the same machine within a pair.
This suggests two ways to greatly reduce spam in our wiki.
One, implement a CAPTCHA on edit, much as the blog does. Possibly allow established users to skip the CAPTCHA stage, which might be advisable if there is any spam to fight at all.
Two, ban by IP. There's a good chance that banning a single spammer by IP would close more than one spammer account.
Two others would be helpful, in my opinion.
Three, allow a shortlist of trusted users to cull accounts based on abuse. Possibly require voting to alleviate the possibility of one supereditor getting a god complex.
Four, norobots, and nofollow.
Edit. Typo.
Edited 1 time(s). Last edit at 02/17/2008 02:06PM by Sean Roach.
Could we use just use Sourceforge? Or preferably install Savane, GForge or LibreSource on our own server.
It strikes me that the development of 'products' in the object library would be best done in the same way as collaborative software development, with versioning, forks, history etc. which these pieces of software are built from the ground up to achieve. The only issue is how we translate the metaphors of software development onto the development of descriptions of physical objects.
If you need any help with this, I do backend servery stuff myself on occasion and would be more than happy to lend a hand.
Sourceforge seems to be getting rather old, slow and clunky nowadays. I also tend to associate it with huge numbers of dead projects, but that's just a personal negative view
I've not heard of Savane or LibreSource before, but GForge hasn't ever looked particularly enticing to me. Perhaps having seen various places use it, badly, has not made the good impression that it deserves.
I'm personally in favour of something wiki like, for sheer flexibility, but with a much less fluffy, friendly, everyone-can-edit permissions model. The core parts of the site should be locked down, and access would be granted after you've proven yourself elsewhere, but creating or editting content in the open areas of the site, or whatever.
On the face of it, a CMS *looks* like a good idea, but having spent some time hacking around with them, let me just say 'don't do it!'. Most off-the-shelf CMS packages are oriented towards blog/news type places, and using them for other tasks is rather tortuous (feel free to browse Drupal's documentation, for example). -- edit... plone looks somewhat more grownup than most CMSes out there, at least.
I'll second the above suggestion of not having a rigid hierarchial file store or search system. A good tagging system is a very valuable thing indeed, and can easily be organised into canned indices if needs be.
Something clever and custom might just do the trick, but that could easily end up as a project in itself.
I am soon to be getting some nice little servers for co-location, and I could perhaps be persuaded to provide access to them, unless more sensible facilities are available elsewhere. This would at least give the possibility of more interesting hosting platforms that are tricky or expensive to come by elsewhere.
Edited 1 time(s). Last edit at 02/18/2008 01:22PM by Ru.
I am in favour of something clever and custom being built, as I feel that getting the object library done well is absolutly vital to the success of this project. It is the first port of call to anyone with access to a repRap or repStrap and so should not only be functional, but also elegant.
It is all very well giving people access to a machine that can print a vast array of what they can imagine and then draw in three dimensions, however the majority of people currently lack the basic skills required to translate their imagination into a 3d file format and perhaps 25% of the population lack the ability to conciously think visually at all - [en.wikipedia.org]
This means that if the repRap is going to be a tool that is to be of use to more than just a rarefied subclass of engineers and artists, then it must have a very usable system for the distribution of 'products' to those who would otherwise be crowding our high streets for the latest fashion or gizmo.
It must also have an evolutionary (and I mean this in the actual meaning of the word) system of 'product development' that is not only usable by technicians, but can be contributed to by everyone from children through to the elderly, whether through commentary, sketches, technical drawings, photographs, computer models or various formats of directly printable data.
Thirdly, it must also have a rock solid support network built into it, so that people who need help in using a 'product' have a wealth of experience to draw on, both from other users and where possible, the originators of the object in question.
The current attempt at making an object library has, I have noticed, a very flat structure. This is fine when there are a few hundred objects, however it will start to break down after it gets much past a thousand and would be totally unusable should there be millions of objects available. Ideally, we could do with a fractal topology that reflects our use of language to describe the material world.
For example, someone searching for a chair may actually find that what suits them best for sitting upon is an object that was originally created as a table and in finding this table/chair among the objects on offer, should be able to link it back to the chair definition, so that others may find the same use. The more people who then reinforce this crosslinking will add to the chairyness of the thing created as a table, and if this table/chair is actually a very good chair then eventually it may appear most strongly in the chair area of the site, with only a faint whisper left of the creator's original tablelike intent.
However, while this navigational structure of descriptive language should be allowed to flow according to the whims of those using the system, there should also be a mechanism for the direct location of a particular object, no matter where it has found itself conceptually placed. To this end a unique identifier should be made for every version of every object on the system.
Ideally this should also be printed whenever the object is, either as a seperate tag, or if possible into the fabric of the object itself in a human readable form. This will allow anyone who has printed an object from the library to immediately access the help, history and genesis for the item that they possess.
-------------------
In the meantime tho, I would agree with getting a wiki thrown up with a public area that is turing test protected [ *edit* - just found out that this is what the reCAPTCHA references were further up the page.. looks really cool, lets get those books digitised ], plus administrators and user accounts for the live project/object areas, cos that could be done very quickly and would be reasonably functional for our needs in the immediate to near future
Edited 1 time(s). Last edit at 02/18/2008 04:28PM by deadgenome.
I can do this... I had started designing something similar a while ago. I even bought a domain for it: protoplot.com (Currently nothing there)
I'm a *AMP guy, and I have some uber-powerful supertools that I can use to get a functional version of the site up where people can play with it. So, what I /am/ is a web guru. What I /am not/ is a prototyping guru. I've mostly been paying attention to RepRap because I think the underlying pholosophy is awesome.
I think the design was something like this:
There are "parts"... Which are basically single objects (versioned). Gears, etc.
There are "materials" which are things you buy (Threaded rods, sensors, chips)
There are "assemblies" which are made of multiple parts or other assemblies, plus (wiki-like) instructions (versioned).
There are "projects" which have multiple assemblies.
I'd also thought about a review systm that grades:
End product quality
Build difficulty
????
I hasn't thought too far into the community aspect, other than basic login/pm/my profile stuff.
It would be nice to see if we could work in Subversion or a similar real-deal versioning tool.
Also, one of the things I had not worked out is this:
If you have part A version 1 used by assembly B...
When part A gets moved to version 2, how do we want to handle assembly B? Auto-upgrading to version 2 seems dangerous. Notify the project owner?
Wiki with versioning. No open areas except for the forums and profiles. Library editing by invite only. Known entities can request and be invited.
Subdivide the wiki's library entries into project spaces so that registered users can start a new project and add objects/sub-assemblies to it from a general list of primitives or from entire other projects.
Add in a myspace style forum for developers and wanna-be developers. This can also double as a measure of their authenticity as a user before being added to the permissions of the library space.
Custom, obviously. Easy, not at all. Worth it, most definitely.
I have a heap of stuff I'd like to discuss on this.
I do think that there is a real need for an interim solution that can be created with existing 'off the shelf' tek.
However there should be a working group that takes on the creation of a custom object library as a not seperate but distinct project.
If we want the repRap to succeed then it is vital that the object library be done exceptionally well. All posts I have made so far on this subject have not been, in my true opinion, 'THE WAY I THINK IT MUST BE DONE', as that is something which I am far from sure of yet.
So, I may contradict myself regularly on this subject as my thoughts on this matter progress. The one thing however that I am sure of is that we make damn sure to build this thing well and learn from the mistakes of those who have come before.
I would therefore suggest that a development team be setup for this purpose with clear goals and which should work from whichever software development site best suits our current needs. And the first matter of business should be working out clearly what those goals may be. (or possibly which site to work from)
*note* I am not necessarily suggesting by that I should be a core member of this part of the project... Historically, if I have been the best organised member of any particular group, then that group is doomed... although I have been getting better at that recently
... i can assist with building/construction or conversion of 3D-objects - in the past i made some VRML-3D-worlds for exhibitions and presentations and for our microtec-fairs or customer-purposes i regularly designed and constructed the objects for 3D-fabbing or visual presentation/animation in raytraced scenarios.
For construction, animations and 'photorealistic' rendering i used/use 3D-Studio or Lightwave (sometimes PovRay too ) and it's no problem to output in STL-format for the reprap-object-library ...
I used to swap 3d modelling for beer and smoke when I was at university to help out the fourth year students... there was one of the degree shows that was promoted almost entirely with my images, though the Uni didn't know that, they had just gone through the forth year subissions and picked the pictures they liked...
I've also got a realtime 3d engine with texturing that I've written in flash with an x3d file parser so I can draw stuff up in all the major commercial 3d packages, export to x3d and the flash then can pick up the x3d and the texture file referenced by the x3d, then it draws the object complete with the UV texture mapping.
It probably isn't high polygon enough to use for displaying complex objects, but we could probably do with something similar to allow people to see objects in the library.. perhaps the viewpoint 3d plugin or something...
While we're hashing this out, I'll put up a newer version of the mediawiki at objects.reprap.org this weekend, and then spend a while loading the RepRap printable parts onto it.
I'll think I'll spend a little while longer spam-proofing it this time.
I'm still not convinced that a mediawiki will meet our needs, longterm.
I would be more than willing to assist in this if additional manpower is needed, but unlike some who have posted so far I don't have any particular skill set that would let me be a driver.
Also I would put my vote behind those suggesting something along the lines of GForge/Savana. My experience has been that wikis are particularly bad for hosting and versioning things other than text. So I don't think it would work well for storing, versioning and accessing the various types of files that would be generated.
I think that a GForge like system would do a good job at providing the needs of scenarios B (structured files) and C (personal and community communication spaces) but not so good at A (random dump location).
The hard part is that things like GForge are project centric. It would be unwieldy to have to create a project for each and every model. I imagine making it contributor centric by making each contributor a project. Then the contributor would then host their one or more models under their name. However this will leaves the problem of finding models by criteria other than the contributor.
Tags perhaps?
On that note, is there something flickr-ish that could house files other than images? That might facilitate all the scenarios. However I don't think it allows for any kind of versioning.
I'm pretty convinced that a mediawiki doesn't meet our eventual needs, but it will definitely do for the moment and will buy us some time to build that which does...
I also see no problem with every object being a separately referenced 'project', it just depends on what we mean by that. Personally I would make every different version of every different part of every different object be the basic unit of organisation, with high-level concepts such as projects being nothing more than a possible perspective to take on this universe of virtual forms.
I love blakes modular approach, but I would ditch the hierarchy of differentiation. Even a single gear is made by layers that can be arranged in an infinite variety of ways.
If we organise by creator, this breaks down when we have several creators for a given object. Complex 'products' are rarely the offspring of a single intellect, we need to build in the opportunity and tools for collaborative thinking right from the start.
The important thing is in not making the walls too high between projects. People should be encouraged to borrow and plaigarism should not be a dirty word, after all it is often the most honest form of complement.
One thing we should definitely be doing is finding mechanisms that allow one person to develop a versioned thread that charts the ongoing development of THEIR baby and does not hurt the egotistical attachment to it, for this provides for many people the creative environment in which they flourish and produces some really wonderful things.
But at the same time, the other thing we should definitely be doing, is providing the methods by which people and groups can create cross threads, referencing whichever versions of other peoples creations or parts of creations suit their needs best, without being subjected to the fact that the 'creator' of the part that they utilised has just 'upgraded' the design to version n.x and in so doing has removed the very feature that they need. However they may like to be informed if someone on the system has created any upgrades to that part, but then again, they may not.
Remember that progress is not absolute, but rather is merely an artifact of perspective. This is especially true in design and is always true in fashion.
I could definitely help with the design (that being my strong arm) and prototyping (using Flash), but I don't know enough LAMP stuff to implement a solid version. I think it could easily be done in Flash/Flex (and a mobile version using Flash Lite), but I'm not sure how well received that would be. While Flash has over 98% penetration (of all net users in the world) and such a thing could work wonderfully with the platform, a Rich Internet App approach might be too "heavy" for such a problem. Hard to explain, but we can talk about that later.
But before we even begin to think about the design of the new site, we need a better collaboration environment in which to discuss the many aspects of it. A single thread in a primitive forum really isn't good enough to hold lively discussions for such a project. We should at least create a new sub-forum dedicated to this, or install a full-featured forum elsewhere (I'd like to suggest SMF). If bandwidth or space or whatever is lacking, I can provide hosting at my own space without a problem.
So as soon as a better space to discuss this is provided, we can get busy. The proposed site is just too complicated to discuss in a single thread (for there are many parts to discuss simultaneously).
I've been learning AutoDesk Inventor. I think this has some good ideas for us. Organize around projects. A project can be a single item or multiple items. Each project is made up of assemblies of parts. Each part--the most basic unit in this scheme--gets a version. Save it to the site and a version stamp gets applied perforce. Projects can have contributors but are required to allow cross project stealing of parts or assemblies or the entire project.
I think an object library is pretty important. Since the current wiki is locked down until the spam is fixed, how about a 'quick and dirty' interim solution? that way we can capture all the built objects and transfer to a 'proper' library later on.
I suggest we use a new section of this forum - 'object library'. Add a single 'sticky' post with some rules/guidance (as a suggestion)
1) Please create one new thread for one reprap object (e.g. shoes, coat hook)
2) The initial post should contain
a) a brief description of the object
b) the STL file
c) the 3d files (if available) - blender, etc.
d) a picture of the finished item (if available)
e) a review of the item (did it work? - if possible)
3) Please post improvements, suggestions, new copy photos, etc in the original thread.
At least we can record all the stuff people are making at the moment.
We can then transfer to the new software when available.
As to the 'final solution', it sounds to me a bit like a forum (to control posting of new objects and record a formal history) with a wiki page for each object. (or a wiki with restricted users - a sensible captcha system)
I like your discussion about this topic,the overall management of the library and its concerning problems. I think an object library is pretty important.
If there is a webserver that could be used to store the stl-files,
then there is more than just one way to do it.
You may can use some regulare forumsoftware (like vBulletin or phpbb)
and just attach the stl to the post (Proper Topics required though ) or the added download section,
that way one could discuss the files, show renderings, printouts and so on.
Or you could "mis"use a video or music library, but you'll need to add a filehosting script that may cause security issues.
And it's not as easy to add a new item as with forum software.
You can of course use nothing but a fullfeature filehosting script too,
but that way it's not as easy to sort the items by intention.
There are many pros and cons for each.
The BEST thing would be, to write a script yourself, that way it'll have all pros and less cons
If you need help with programming some php or such, just let me know, maybe I can help.
Do you guys like the general feel, or should we start looking at a video or music library-like approach?
I like the fact that since it is a wiki, we can use it for development of future RepRap versions, and other projects, like a laser cutter:
[objects.reprap.org]
Right now, the biggest downside is that it doesn't have very many files on it.
[objects.reprap.org]
On that note, any suggestions how we can get a lot more contributions?
SebastienBailard Wrote:
-------------------------------------------------------
> Right now, what do you think of what's at
> [objects.reprap.org] ?
>
> Do you guys like the general feel, or should we
> start looking at a video or music library-like
> approach?
>
> I like the fact that since it is a wiki, we can
> use it for development of future RepRap versions,
> and other projects, like a laser cutter:
> [objects.reprap.org]
>
> Right now, the biggest downside is that it doesn't
> have very many files on it.
> [objects.reprap.org]
>
> On that note, any suggestions how we can get a lot
> more contributions?
I don't think a wiki alone is the best solution because it requires anyone wishing to upload a model to learn the formatting guidelines for themselves, because they need to design their own wiki page.
Perhaps a wiki page could be embedded within a more conventional music-type library. I'm envisioning an uploading script that prompts the uploader for basic details on their model (dimensions, volume, suggested printing material, short description, picture, etc) as well as the STL files, and any available "source" files such as .AoIs. It then generates a base MediaWiki page describing the object. I would presume the object uploader would be able to decide who is allowed to edit the page - himself, a select group or the general public. He retains admin privileges.
Basically, starting with a wiki I think you need to add (1) a "conventional" model database with basic stats for searching and sorting, (2) a user-friendly (think google) search-based user interface that returns relevant results, and (3) a wiki-setup-automation tool that takes all of the hard work out of formatting the initial MediaWiki page for a new model. A template may work too, but I think the form is more foolproof.
SebastienBailard Wrote:
-------------------------------------------------------
> Right now, what do you think of what's at
> [objects.reprap.org] ?
>
It may be me, I'm not a fan of wikis.
Not for stuff like this.
> Do you guys like the general feel, or should we
> start looking at a video or music library-like
> approach?
>
I would say so, it's hard to browse wikis without knowing what you're really looking for.
if it's more like a video database one could browse to get some ideas,
scroll a "genre" and then select an "artist" something equivalent would be nice.
(that's why I like forumbased systems; generally there's more infor for the file than just the stl and a picture)
> I like the fact that since it is a wiki, we can
> use it for development of future RepRap versions,
> and other projects, like a laser cutter:
> [objects.reprap.org]
>
True, but we already have a wiki for that kind of stuff, don't we?
the complete homepage is nothing but a wiki
> Right now, the biggest downside is that it doesn't
> have very many files on it.
> [objects.reprap.org]
>
That's because there aren't many RepRaps out there, so there are not too many people already printing stuff other than the darvon parts.
As soon as the number of active Printers increase I think the number of files will too.
> On that note, any suggestions how we can get a lot
> more contributions?
First: Start selling the extruder parts in the store again, so the hardest part of the darvin can be bought (see above)
Browse the Forum for already uploaded stls and copy those to the library
(not many people know the library at all)
Make it as easy as possible for registered users to upload to the wiki
It's somewhat confusing if I have to register myself numerous times within one website to do different stuff.
(another pro for a strikt forumbased library... that almost looks like a commercial; I'm sorry)
Oh btw:
I find it somewhat confusing with all the different folders and fileversions at sourceforge.
( I'm not as engaged in sourforge projects, so I'm not familiar with how they deal with different stuff)
It'd be nice to have the latest version of every stl listed in the library too.
That'd be much easier I guess. (at least for those like me )
And to get back where I started:
I like to see more than just what the originator of the file thinks about it.
I'd love to see improvements on that file, photographs of "what happened when I printed" or just reading a "thank you that's exactly what I needed"
If the uploader gets some feedback on his work, it's just more fun.
All that is impossible with a wiki (unfortunately)
Im a 3d artist amd would be willing to post some stuff that i make, also as a file location i would recommend turbosquid.com (a 3d warehouse site) you guys could create an account and upload a library of stl's and such as free assets which would allow you thumbnails and such as well. (and unlimited storage there is a guy with 9000 texture images on the site taking up oodles of space so they dont care at all about your space consumption),
Honestly i know of a few people on turbosquid that would be interested in the reprap movement. (and would make a hat load of models just for the heck of it)
I plan on building a reprap some day, but all my money goes into software right now as well as my college so im straped for cash (which means that by the time i can afford it the third gen model will be out)
As of right now i could give you guys the stl of my plasma genorator and even include a base like a warmachine figurine
I think you all should head down to www.3dcontentcentral.com if you want files.they have almost anything you could ever need in a variety of file formats.Seriously, almost anything!A lot of manufacturers also post their solid files there for use by engineers, students,etc.
P.S, I love the reprap project and I look forward to making my own one day. TTFN, Joe.
Hi people.
I am very excited about this post im writing now.
I read this topic back when it was started and already then I had an idea for a site like that. But reading it here, was what really got me started.
I didnt post anything when i started it because i wasn't sure i was up for the task.
But, without further ado, here it is: www.reprap-models.com
The database is at the time of posting completly empty but in the days and weeks to come, we will be posting different models we deem helpful or necessary to the reprap community.
For two reasons i will not upload other people's models first of all because i have no right to and even opensource, free to share models, i will not upload because of the way the website is built up.
Every item has a page and is linked to the account of the creator.
Therefore if you want something specific, like the actual reprap parts,
please, please ask the creators of the objects.
I apologize for multi posting, but i wanted it both here, in this thread where i got the motivation, but also in its own topic.
Edited 1 time(s). Last edit at 11/07/2008 12:48PM by logan.
], plus administrators and user accounts for the live project/object areas, cos that could be done very quickly and would be reasonably functional for our needs in the immediate to near future 
)