Ticket #425 (new story)

Opened 13 years ago

Last modified 13 years ago

Material amortization

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

Description

Note: The most of curicula has natural amortization (ex.: a material about IT, which is modern in nowdays, will be old in the next year). The ranking and searching system should be deal with it. Toolbox could be use a "usage factor" for it. This usage factor is a hidden property of a material. If anybody use the material, the usage factor is increasing (pupular materials will have high usage factor). This usage factor should be decreasing with amortization. Amortization is an other property of the material (ex.: fast amortization [ex.: IT technology]: -10% from usage factor / week, slow amortization [ex.: phisics] -1%/week). When a teacher searches a material, the materials, which has high usage factor should be the begining of the search list.

Stroy: Teacher searches IT releated materials. He has found good materials, which are modern, and usable. The teacher is happy, and he thinks Toolbox has a very good ranking system.

Change History

comment:1 Changed 13 years ago by teemu

This is a good idea, but I would approach this more from the “Google’s way” of bringing relevant stuff to the top in the search result.

When putting in order the search results we should look at least the following factors:

  • Use of the material based on the incoming links
  • Activity in the editing the material
  • Something else?

Based on this we may even show in the search result page with icons or something the most “used materials” and the most “active projects”.

I like the idea of a amortization index, too, but I am not sure if anyone (we or the users) are able to give these indexes. Example from IT: In 1980 when computers were not very powerful it was important to be able to write simple and small software. When computers got more powerful it was not anymore considered important. In 2000+ with smart mobile phones – that are not that powerful – the skills to write small and simple software is valuable again. So things are changing but it looks for me that there are more circles than flows. So, one should be able to change the index in different times, but as said already this all makes it very difficult to define the index. It becomes very fussy and relative. What i high for me is not neseccary high for someone else.

comment:2 Changed 13 years ago by teemu

Sorry - took another look of the story by Laszlo and hopefully got it now right. The amortization index is given automatically based on the use of the material. This way my first proposal of the raiting content based on their popularity is basically exactly the same feature.

comment:3 Changed 13 years ago by laszlo

Teemu!

Could you write me your tickets number? I could compare the two things.

comment:4 Changed 13 years ago by hans

I was thinking about this problem during lunchtime and I am not sure that we are going to have problems with outdated learning materials.

School textbooks get outdated, when something big and unexpected happens such as 9/11 or Asian tsunami. I case of LeMill there is usually a main learning material on each topic and some copies/translations. I am sure that the main material will be updated quite quickly. In a working community there is this I-want-to-be-the-first-one-to-update-this-material feeling among the members.

Step-by-step materials about Microsoft Office XP will get outdated when new Office version will come out, but in this case tagging helps much better. Learning materials about Office software will probably have tags like office97, office2000, officexp.

I support the criterias that Teemu mentioned (incoming links, activity in editing).

comment:5 Changed 13 years ago by laszlo

OK. What you told is true, but the system possibly will contain many unmaintaned materials, and the system should deal with this. I think, the bigest part of materials will be unmaintaned. Many teacher will try the system, create materials, and after that, they will forget these. In an ideal world, the Calibrate repository would be a big nice Library with usefull knowledge, but in the real, this repository will be polluted, very polluted (like the Internet), and if we don't help the teachers to find the good materials, they won't use the system. For this reason I think, the system should support features like measurement of usage of materials, amortization of materials, ranking, contant based searching, and searching in trusted groups.

comment:6 Changed 13 years ago by hans

I wrote a long comment but lost it :(

Shortly:

#424 and #425 are both about ranking and ordering the search results.

For good search we need:

  • collect data about usage of learning resources
  • create search algoritm and tune it when needed

Two people in KU Leuven are actually doing research on that field. They said that there is no good open source search algoritm for learning objects that we can take into use. However, they have done research on collecting the empirical metadata about the usage of learning objects: http://ariadne.cs.kuleuven.ac.be/empirical/

The data that they collect is not 100% relevant for us (we are not going to have 1...5 rating), but we can learn something from that. It is very possible, that in few years their CAM framework will become similar widely acceted and used specification as SQI is now.

I will create a new story about collecting usage data. Then we can have that discussion under one ticket.

comment:7 Changed 13 years ago by laszlo

OK, but usage data collection and material amortization is two different thing!

Note: See TracTickets for help on using tickets.