So here is what I propose. I say we carry on with the current path of pseudo 
merging device and OS detection. I think we all agree its not optimal, so lets 
do our best to keep it sane (and keep our sanity). There is definitely value in 
detecting operating systems and it wont take too much work. This is our version 
1.X path.

As for DeviceMap 2.0, we should plan on separating and partitioning 
classification between devices, operating systems, and browsers. These are all 
separate classifications with their own attributes which can be detected 
independent of each other. This is the proper way. It may require overhauling 
our version 1.X work, but thats why software has 2.0 releases :)

Sound good?


________________________________
 From: eberhard speer jr. <[email protected]>
To: "'[email protected]'" <[email protected]> 
Sent: Thursday, July 31, 2014 8:53 PM
Subject: Re: generic desktops, crawler, and generic android
 

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I think the main problem is this :

If you look at the W3C definition of DDR Vocabularies [1] and the
definition of the PropertyRef Interfase [2], it's clear that all
Properties must be accompanied by their Aspect or have a default
Aspect defined.

The W3C DDR 'defines' 2 Aspects [3] : device and webBrowser (but
mentions others [1])

So, the property "Vendor" has both aspects and when not specified the
default aspect is "device".
This means :

Vendor = Nokia
- -> refers to Vendor in it's default aspect : the device vendor

and

Aspect = webBrowser
Vendor = Nokia
- -> refers to Vendor in it's webBrowser aspect : the webBrowser vedor

ODDR, added the aspect "operativeSystem" [4]

*But*, after a while all properties ODDR added only had one aspect :
device.
Over time the whole 'Aspect' thing was 'forgotten' and *every*
Property defaulted to 'device'.
Having 'lost' the Aspect...err, aspect -- also in their interfaces
[IBuildObject] -- ODDR ended up with properties like "os_version".

So, the whole Aspect thing got mangled !

Our main question now, I think is : do we follow [resurrect] the
Aspect standard or go 'the other way'...

Just off the top of my head : with a notation :

Aspect.Property=Value

instead of today's

Property=Value

I'd say follow the Standard !

esjr


[1] http://www.w3.org/TR/DDR-Simple-API/#sec-vocabularies
[2] http://www.w3.org/TR/DDR-Simple-API/#sec-PropertyRef
[3] coreVocabulary.xml
[4] oddrVocabulary.xml
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJT2uUmAAoJEOxywXcFLKYcGOcH/jB8xR2IEkzFLGUNcv8WRWr0
55r7iJE7w8rmbXdzYC3cMfbXRlC1fL+4pvf/8djfDkrTAgLl02FMlBVOnooxrs1z
IYJzT2Ldrh3oPDfvY2ecPTkPmlFRvtG2dxEng/FcuciwDNOe/4VGvPLEIQ5qjRtx
654G3GVNIWrm6ypNKG3KovOEJxMB6cGzXdrAasF2r95tnKIUQEdNS2I69BIGQvTr
p8m+rxKcZPEAvm0Pm5Gu2HFrJrxgUjJktFVNW4ZLIsKnlTfLe3PfObJTAvXNcbh9
oxgAP6xlXe2UZo/mG+LHUthdVi1hhxwfd87Q39oapq0VtKE187HVaTbVkuzQRkc=
=b7vZ
-----END PGP SIGNATURE-----

Reply via email to