RE: does file content deserve its own class?

 

> Carlos Iglesias schrieb:
> > On the other hand, and I am just wondering, if the file is 
> not publicly available, do people really want to share the file? 
> 
> There are privacy levels between internal only and available 
> to the public. E.g. consider a trusted external evaluation 
> team that evaluates a non-public resource.

Sure, I'm just wondering about the possible use cases we have when the file is not publicly available. My guess:

A - local files (file://...)

It usually happens when teams work on templates or when there are really small teams that produce "handmade" websites (usually static), not sure if EARL has something to do with this case.

B - local testing servers (localhost://...)

Mainly again in small projects, but I think that they have different characteristics (usually dynamic web sites). Probably this use case is more relevant for EARL that the previous one.

C - internal testing servers (156.89.98.3://..)

Used under development before production stage. This could match with the trusted external evaluation team you commented. EARL should be able to cover this uses case.

So my impression is that there are enough differences to justify the creation of a new specialized class and we need something more generic that FileContent to allow us the acommodation of all previous use cases.

> >>  * File URI can not be assumed to be unique and so a property to 
> >> store the file name is needed as a label for the file 
> contents (see 
> >> also related comment by Charles [1]).
> > 
> > The problem with this approach is that we can easily find 
> two or more 
> > "index.html" file names for example,
> 
> I think, it was a file name plus the path.

Well, there could be some "security" or "privacy" concerns that prevent people to expose the whole path as pointed by Charles before [1]. Anyway neither of them are useful as unique identifiers so we could use whatever of both just as labels.

> > and then the only way to see the differences is looking at 
> the file content.
> 
> That's why it is just a label, not an identifier.

And that's why I think we need some property (other than the file content, which is too hard to manage as identifier) that works as a identifier. My justification for this is basically that if you want to share the results (for example with an external evaluation team) you need an unique identifier (for example, is supposed that the external team is going to manage more projects).

Finally, another issue that have just came to my mind is:

How should we manage that content which is avalaible on the web, and thus we can use the WebContent class, but has restricted access (e.g. proxy or user and password), and thus fits more in the "NonPublicContent" concept? 

Should we redefine the WebContent class as "Information that is PUBLICLY AVALAIBLE on the World Wide Web" and then create a "NonPublicContent" that includes file content, password protected content, etc.?


[1] - [http://lists.w3.org/Archives/Public/public-wai-ert/2006Oct/0010.html]


Regards,
 CI.

 
--------------------------------------

Carlos Iglesias

CTIC Foundation
Science and Technology Park of Gijón
33203 - Gijón, Asturias, Spain 

phone: +34 984291212
fax: +34 984390612
email: carlos.iglesias@fundacionctic.org
URL: http://www.fundacionctic.org

Received on Wednesday, 18 October 2006 10:51:37 UTC