Re: Latest version of glossary
Date: Fri, 24 Feb 2006 22:25:10 +0100
Message-ID: <43ff7973$0$24373$e4fe514c_at_news.xs4all.nl>
Alexandr Savinov wrote:
> dawn schrieb:
>
>> ...entity: a thing of interest
>
> Although it is an official definition it has the following problems:
> - it is too general and hence useless
> - it is equivalent to the term "thing", which is used to refer to anything
> - entity != dimension, entity != attribute, entity != reference
The purpose of the glossary is /not/ to provide formally correct definitons of terms. There are plenty of sources for those.
Typical glossary entries just weed out real misunderstandings. 'Entity' is rather hard to define correctly, but the current definition does a good job avoiding misunderstandings. Dawns addition about XML-entities does help this goal of the glossary.
Having said this: IF there is a formal definition which also helps avoiding long misunderstanding driven threads , I'ld be happy to link to it - this goes for all other terms as well.
> So entity is much more concrete term than simply a thing of interest. I
> would define it as follows:
>
> entity is a thing of interest which has properties, that is, can be
> characterized by one or more other (primitive or non-primitive)
> entities, and has an associated identifier.
How can a thing be of interest and not have properties? I haven't ever seen this misunderstood.
> In other words, without an identifier a thing is NOT an entity. Without
> properties it is also NOT an entity.
I really like the identity idea. However, adding it would not serve the glossary's goal, IMO. On the contrary.
> I do not insist on this definition of course but I simply think that
> having too general definitions is not very useful. It will be simply
> misleading in discussions.
>
[snip dimension stuff: not copy & paste material]
> flat = the absence of hierarchy (multiple levels of details)
Copied. Thanks. Received on Fri Feb 24 2006 - 22:25:10 CET