[Data-modeling] from software releases to operating systems...

Wed May 7 18:49:51 UTC 2008

sounds good, Mark.

Robert, yes the expected type "Operating System" for a software  
release/version is problematic. We need to sort out what we mean by  
Operating System, Operating System Release/Version, Operating System  
Series, Computing Architecture (or Computing Hardware, as Tim  
suggested), etc.  Arguably a software release/version could have  
multiple properties pointing to instances of each of these target types.

e.g. Pro Tools LE 6.9.2 * PPC Mac * Mac OS Tiger (10.4)

an actual example of a software release that will not run unless all  
of the following requirements are met:

Computing Hardware: PowerPC Mac
Operating System Series: Max OS X Tiger

and I've seen other releases of this software that actually require a  
minimum version of software e.g.

Mac OS 10.4.9

and some of those same Pro Tools versions actually prohibit a version  
of Tiger that is too far advanced (!) Welcome to the world of audio/ 
video production software :) ugh, I may be willing to draw the line at  
this last level of detail.

I think we should scope out the types mentioned above and create a  
property for each that can be specified for a given software release/ 
version.

Brendan

On May 6, 2008, at 7:27 PM, Mark Verber wrote:

> I would agree that the current OS model needs to be updated.  I  
> think that
> is reasonable to distinguish between the "name of the OS family", the
> release/version number, and the release series name.  But I don't  
> think
> that's the end of it.  I think we would also want to minimally capture
> architecture types supported, kernel, and packaging technology.  I  
> have
> notes on this in a lab notebook which is slightly buried.  When have  
> have a
> spare moment I will try to unearth my notes.
>
> --Mark
>
>
>
> -----Original Message-----
> From: data-modeling-bounces at freebase.com
> [mailto:data-modeling-bounces at freebase.com] On Behalf Of Robert Cook
> Sent: Tuesday, May 06, 2008 7:22 PM
> To: Freebase data modeling mailing list
> Subject: Re: [Data-modeling] from software releases to operating  
> systems...
>
> I may be missing some of the subtleties in this model, but I'd like to
> double check.  If I say a piece of software runs on a particular
> operating system, am I able to say something general like "Windows",
> something more specific like "Windows NT" or something highly precise
> like "Windows 2000 Server Edition SP1"?  That is, are all of these the
> same expected type in your model and if so what is that type?
>
> R
>
> On May 6, 2008, at 5:44 PM, brendan wrote:
>
>> Hi,
>>
>> I'd like to broach the discussion of operating systems and their
>> releases/versions. To recap my proposal for software and its  
>> versions/
>> releases (from last week):
>>
>> "Software" has "Software Releases" which are unique and identifiable
>> by a version label and a release date and one or more Operating
>> Systems (more on this last part, later).
>>
>> e.g. the software "Mozilla Firefox" has
>>
>> "Firefox 2.0.0.14 for Mac OS X"  * 2.0.0.14 * April 16, 2008 * Mac
>> OS X
>>
>> "Software" also has "Software Release Series" which are more flexible
>> in use. This is any series of software releases that are deemed to be
>> related in some significant way.
>>
>> e.g. the software "Mozilla Firefox" has
>>
>> "Firefox 2"
>>
>> which would, like software, have a list of all the software versions/
>> releases; the versions associated with the series.  Yes, it falls a
>> little short on the normalization front but I think it's quite a bit
>> more practical than trying to do some kind of hierarchical thing  
>> here.
>>
>> So, on to operating systems and their version/releases. How it's
>> modeled now:
>>
>> type: /computers/operating_system
>>
>> property: parent_os                      expected_type:
>> operating_system
>> property: includes_os_versions expected_type: operating_system
>> property: version_number           expected_type: Text
>>
>> I'm not fond of this schema.  I'd like to deprecate it and model
>> operating systems similar to the way I modeled software.
>>
>> "Operating System" (e.g. "Mac OS X")
>> "Operating System Release/Version" (e.g. "Mac OS 10.5.2")
>> "Operating System Release Series" (e.g. "Mac OS X Leopard")
>>
>> The main difference is os release versions don't have a supported OS/
>> Platform like software but, rather, a supported "Computer
>> Architecture" which refers to, well, I'm not sure. Should it be a  
>> list
>> of CPU series supported? I don't think that's such a good idea.  I'm
>> afraid we may need another type here: "Computing Architecture"  e.g.
>> x86, x86-64, PowerPC, XScale, StrongARM.  I think this would cover
>> most use cases.
>>
>> Now back to Software Releases.  I said they are characterized by a
>> version, a release date and an operating system.  That last property
>> is a little fuzzy.  I mean, some software releases are compatible  
>> with
>> an operating system but only a minimum version of an operating
>> system.  Worse, some software releases are only compatible with a
>> particular operating system for a particular computing architecture
>> (e.g. stuff that only worked on powerpc macs running OS X)
>>
>> Before I dive into this last problem I'd like to get some feedback on
>> the general direction I'm heading....
>>
>> thanks,
>>
>> Brendan
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> Data-modeling mailing list
>> Data-modeling at freebase.com
>> http://lists.freebase.com/mailman/listinfo/data-modeling
>
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling
>
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling