Ticket #426 (closed story: duplicate)

Opened 14 years ago

Last modified 14 years ago

Content based material searching

Reported by: laszlo Owned by: anonymous
Priority: major Milestone:
Component: generic Version:
Keywords: Cc:
Time planned: Time remaining:
Time spent:


Note: In many cases teachers would search about the content. For example search about a part of a poem, or search a material, which has bad metadata part. Content based searchning is an easy thing with text based content (like text and HTML documents), but not an easy thing if we want to search in a binary document (PS, PDF, DOC, PPT, etc.). Toolbox gives the possibility to upload materials in these formats. For this reason, we would develop a google like mechanism to searching in these documents. If the teacher upload a document, and the system know the format of this, it is convert it to text format (formating, and images are not important, we would use it only for searching about text content). If the system cannot convert the uploaded file, it writes a warning message to teacher, that the content based searching won't work on this document.

Story: A teacher searches material about a part of a poem. She has found a good PDF, which is contain the full poem. The teacher is happy, because she has found what she was searching for.

Change History

comment:1 Changed 14 years ago by tarmo

This sounds pretty good to me - full text searching is an important feature. And Plone already can convert many document types into text, so this should be implemented at some point. Not a high priority yet, though.

comment:2 Changed 14 years ago by hans

Maybe we should be evil and NOT implement this feature? This is a clear sign that we prefer not to have office documents in our system. Uploading Office document (or better PDF) makes sense only for long documents and if it is a good document, then the users will find it anyway (many collections linking to that document, tags, some blog posts, ...).

If we provide good tools for working with binary files, then some users might prefer these to our authoring templates. Let's leave office users with EUN portal and open-minded people with our online-editable templates.

Uploading a binary material makes sense for things that cannot be done with online templates:

  • flash movies
  • maps, diagrams and other images with multiple layers (can be edited by other users who have software and skills)
  • materials based on excel spreadsheets
  • ...

I don't think we should ban DOC, PPT, etc, but we should not support it in any way.

comment:3 Changed 14 years ago by laszlo

I understand you, but I think it is not a good thinking. As I've written, teachers like their WYSIWYG editors (I understand why you don't like, but they like). For this reason they WILL upload their DOCs, PPTs, etc. (you won't be happy, but they will do this). And after they try to find a good material about Toldi for example. There are 20 very good DOC (which are uploaded by other WYSIWYG fanatic teachers), but this teacher won't have found any dcument. What will he/she think? "It is a stupid system, cause it has nothing about Toldi". He/she will try the system 3 other times, and after these 3 unsuccessful tring, he/she will forget the Toolbox.

comment:4 Changed 14 years ago by hans

Maybe we should define more clearly what kind of content we want. I would comment here David Wiley's RIP post on learning objects:

So whether learning objects are dead or not, I couldn’t say. And to some extent, who cares? As long as people are willing to (1) openly share (2) educational materials that will (3) render properly in most web browsers, and they also (4) provide access to the unobfuscated source for the materials (especially for Flash files, Java applets, Photoshop images with many layers, and the like), I certainly don’t care.


I would emphasize the third property of "learning objects": render properly in most web browsers

I am starting to think that we should limit our system to:

  • pieces (images, sounds, videos and animations that render properly in browsers)
  • materials that are based on online editable templates and pieces

Then we have a problem, that in some cases sound, video or flash animation has already so much pedagogical context that it is a material. I suggest that the media file is still a piece and it becomes a material when it is used in a template. Template for audio and video files should suggest users to do a short summary of the content - it is important for accessibility.

All file types that are not in the whitelist will be simply banned (Office documents, Openoffice documents, PDF files, zip files, etc).

Wiley suggest that learning object systems should "provide access to the unobfuscated source for the materials (especially for Flash files, Java applets, Photoshop images with many layers". Therefore we should provide possibility to "add source file" to the pieces. It should be one link in the actions portlet, not a confusing field in the Additional information form. Materials will never have source files, because they are based on templates and pieces.

Simple and beautiful.

comment:5 Changed 14 years ago by laszlo

Simple and beautiful. The only one question is: will it work? If teachers think: ok, this system is better than my PowerPoint?, Word, OpenOffice?, the Toolbox will win the war. But if this system cannot persuade the teachers, thay will think: "I cannot use my beautiful PPT-s, DOC-s, etc., I don't want to use this system". An another feature, what we loose is the SCORM import. Ponder this things before ban these from the system!

But ok, follow this philosophy, teachers will adjudge it is good, or not. But if the templating system come to the front, I think it should be changed to a flexible pluggable way (users can download/search/create new templates). Hardcoded templates are not too useful.

comment:6 Changed 14 years ago by teemu

I agree with Hans and David: "all content should render properly in most web browsers".

The separation of pieces and material with the factor that the first one is made locally (in your client, PC etc.) and the second one is authored with templates online in LeMill makes sense. This will make it much easier for users to understand what is a "piece" and what is a "material".

I also agree with Laszlo that the template system should be designed so that one can design new skins into it. However, I think designing new templates is an expert's job and there is no reason to provide it for the basic user. This means that we do not need an UI for editing the templates but UI to “add” new templates.

However we may - I think - provide the basic user the possibility to move the pieces in the templates from up to down, down to up / left to right, right to left (like the blocks in Moodle.

I think we made a decision. Right?

comment:7 Changed 14 years ago by laszlo

I think, the question is: what is a template? If templates are only layout templates, the possibility of moveing peaces (what Teemu has written) is enough. But in my vision a template should be a full visual template (with backgrounds, etc.), an XHTML page with slots, and users can put peaces to these slots. It would be flexible, and more extensible. What dou you think about it?

comment:8 Changed 14 years ago by hans

  • Status changed from new to closed
  • Resolution set to duplicate

I will close this story because it is not relevant anymore. We do not allow uploading DOC, RTF, ODT, PDF, etc as materials.

Advanced search form will have possibility to search from the body text of templates (#485). So this ticket is a duplicate.

Note: See TracTickets for help on using tickets.