RE: Naming conventions

Hello everyone, 

Andrea is absolutely right that we must avoid name divergence, and that is not the intention of the alternate names. They are not meant to provide a general alias capability. Instead, they are meant to provide versions of the same name that may be syntactically more acceptable for particular environments. So, for example, we have a characteristic in the ontology for the apspect ratio of the pixels on one of the (possibly multiple) displays on a device. The intent is that we might have something like

Ontology name:   hasPixelAspectRatio
Camel case name: pixelAspectRatio  (might be good for a variable in Java)
Hyphenaged name: pixel-aspect-ratio (might be good for variables in scripting)
Title:           Pixel Aspect Ratio (handy for documentation)

In other words, the distinction between the alternate names is meant to be limited to the syntax in the way they are written. It should be clear to a human reader that they all refer to the same characteristic. Also, anyone using any of them can always provide a reference back to the ontology name if they need to define unambiguously what the characteristic actually is. In addition, they are all formally related in the ontology and can be discovered by machine processing, by people who indulge in that kind of thing.

By the way, it would be perfectly reasonable to add, to this list of alternates, the name by which DDWG wants to refer normatively to this characteristic.

I think the ontology name has to follow semantic web best practices, and by the way, I'm not sure they do in the current version. I intend to sort that out before the F2F. I think that means that there will be some decoration of the ontology names (something like the 'has', 'supports' etc. of the current entries). For other syntactic variants, I think its up to DD to choose what it wants to use and how much 'pattern' it wants create around the  names likely to be used in interfaces, examples etc. It seems to me that the decoration in the alternates might need to be different for different purposes, but I could be wrong. As long as its clear that different decorations constitute just syntactic variations, then we should be fine.

I guess the way I see this working, in terms of documents, is that the ontology doc is referenced from the DD core vocabulary doc (top N?) and the DD core vocabulary is referenced from the interface definition docs.

It's worth remembering that there are going to be things in the ontology that are not part of the DD core vocabulary. For those, DD can probably let others agonize over the particular names that get used!

Best wishes
Rhys




-----Original Message-----
From: public-ddwg-request@w3.org [mailto:public-ddwg-request@w3.org] On Behalf Of Andrea Trasatti
Sent: 14 March 2007 21:04
To: public-ddwg@w3.org
Subject: Re: Naming conventions


Once again, I suspect that having aliases might seem of help at first site, but once in the wild and with different companies and developers from different sides of the world with different backgrounds, might lead to confusion.

I would prefer to have a single capability name.

There is a real world example, J2ME Polish, they have a device DB and support aliases. They even added common typos or slight differences in the spelling. I think it's great from a pragmatic point of view, but it doesn't seem the best approach for a group that is trying to determine a standard.

- Andrea


Il giorno 14/mar/07, alle ore 16:29, Rhys Lewis ha scritto:

>
> Hello everyone,
>
> I just wanted to confirm that Rotan is correct when he says that it is 
> possible to have additional associated names for entries in the 
> ontology. Indeed the reason I introduced this capability was precisely 
> to support the notion of different names for the same item in 
> different contexts, as exemplified in Rotan's list.
>
> Naturally, I think it will make life easier if the various names for a 
> particular characterisic are related to one another in a simple way.
>
> Best wishes
> Rhys
>
>
>
>
> -----Original Message-----
> From: public-ddwg-request@w3.org [mailto:public-ddwg- request@w3.org] 
> On Behalf Of Rotan Hanrahan
> Sent: 14 March 2007 13:47
> To: Andrea Trasatti; public-ddwg@w3.org
> Subject: RE: Naming conventions
>
>
> There are multiple name-related issues here, and I thank all the 
> contributors for pointing these out. I think the names break down as 
> follows:
>
> - Names, as used in the Ontology.
> - Names, as used in the Core Vocabulary.
> - Names, as used in the IDL.
> - Names, as translated to programming languages to access the Core 
> Vocab via an IDL-derived interface.
>
> It is only the last case that I think will have the most impact for 
> the success of the DDR. If I understand the suggestion made by Rhys, 
> it may be possible to have additional associated names and a rationale 
> for creating those names, so that software developers will observe a 
> consistent convention in naming. If this is the case, and the burden 
> on the DD group editor(s) can be managed so that this is not an 
> onerous task, I would support this approach.
>
> I don't think the developer community will really care about the names 
> we use in the Ontology, or even the Vocabulary, but when it affects 
> the source code they have to write (and later read, to
> maintain) they will care. I hope this discussion here shows that we 
> care too.
>
> ---Rotan.
>
>

Received on Thursday, 15 March 2007 10:25:20 UTC