Re: Format Tests

Thanks for the clearifying words, David!

Something about your further points:

(Simplicity I) 
I don't see a point in having any specific field first because of the 
following context. The most "natural" data structure used to hold the 
list during runtime is a hashtable. Hashtables have random ordering when 
they are made to serialize. (The end-marker 0xff is not part of the 
content but just a meta-tag on the file!) a) There is no necessity to 
demand extra program logic to bring out the 0x00 field first and the 
rest trailing. b) The "sound and simple" algorithm dealing with field 
lists would take two steps: 1 - reading in the entire list (purely 
formal) and 2 - interpreting fields. This advocates to forget all about 
field ordering.

(Tolerance II)
My idea was to create 3 realms of field types which mean 3 levels of 
"formal and semantical jurisdication", so to say. I fully agree there 
must be a range which is completely secluded for the PWS project. I have 
seen 0x00 - 0x3f fitting. These are 64 fields and should do the job for 
a while. (I don't mind if it is some numbers larger.) Let's call it the 
PWS-HOME (or pws-private) range.

Then there is a second range from 0x40 to 0x7f (or some other range), 
let's call it PWS-APPROVED (or  pws-protected). The fields there are 
formally administered by the PWS project but they are given away to 
approved pws-like other projects for their own and private purposes. 
This means PWS is just holding a list of reservations but does not 
control any semantical aspects of those fields. The advantage for 
"approved" projects is that they enjoy a level of protection against 
random destruction if any "rest-of-the-world" project should be  using 
extra fields.

The third level is WORLD and ranges 0x80 to 0xfe (or some other range). 
It is meant for projects and experiments which never perk a head in the 
discussion lists. This range is not controlled by PWS at all, but a 
recommendation can be given out for applications to perform 
data-marking  in order to verify the property of any read fields.

Regards
- Wolfgang

dk wrote:
> Wolfgang,
>
> Honest - I am really not intending to cause any "unnecessary obstacles".  I
> apologise if my posts appear so - that was not the intent - probably my use
> of the English Language.  I value other people's contributions and ideas.
> The more the better.
>
> Nothing I have said in these posts is fixed - they are only my current
> opinion and could easily change.  Rony is the final arbiter and there is no
> guarantee that he will pick my view, your view or anyone else's; that is why
> I have asked for a "draft specification" for the next and future formats so
> they can be discussed more widely.
>
> I have put my answers in the body of your message below prefixed by ">>>".
>
> Regards,
>
> David
>
>
>   
>   

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/