Re: software version number datatype

From: HansF <News.Hans_at_telus.net>
Date: Thu, 28 Apr 2005 14:24:35 GMT
Message-Id: <pan.2005.04.28.14.26.45.919282@telus.net>

On Thu, 28 Apr 2005 09:18:56 +0100, D Rolfe interested us by writing:

>
>
> HansF wrote:

>> On Wed, 27 Apr 2005 07:36:55 +0100, D Rolfe interested us by writing:
>> 
>> 
>>>Stay away from custom objects.
>> 
>> 
>> Why?
>> 

>
> Look at it in the context of the posters's question - If you accept that
> version numbers are actually more complicated than they look then
> bringing objects into the equation will probably only make things worse.
>
> Modelling software version numbers strikes me as one of those problems
> which looks just fine for the first 5 minutes of whiteboard modelling
> and then turns into a bottomless pit of complexity when you encounter
> real world data. A solution which works perfectly for one manufacturer
> will fall apart when other makers products are brought into the picture.
>
> Software product version numbers are not logical - they are partly
> determined by the marketing department......
>
> All of the above is my humble opinion. If you believe different things
> then that's fine with me.

No argument. Just interested in an expansion to what seemed like a blanket, and potentially myth-related, statement.

Would you use the same argument (about staying away from objects) in context of IP addresses, which are reasonably well defined?

/Hans Received on Thu Apr 28 2005 - 09:24:35 CDT