Re: Objection to change in svn_depth_t component naming (r21067)

[Karl Fogel <kf...@google.com>]

Max Bowsher <ma...@ukf.net> writes:
> Sorry, I wasn't following this thread and didn't comment before - I've
> just noticed that such a change was committed in r21067.
>
> I would like to object on the grounds that the zero/one/infinity
> naming scheme is far more intuitive than the LDAP names.
>
> This isn't just a reflexive action the change on the first time
> encountering the LDAP names - I've run into them before, and found
> them unnecessarily confusing about the precise concepts they refer to.
>
> Whilst "onelevel" is reasonable, neither "base" nor "exact"
> communicate their concept to me as clearly as "depth zero". Ditto for
> "subtree", which makes me think of selecting just a particular subtree
> of nodes 
> From underneath the object it is being applied to, not entire tree 
> rooted at the reference object.

I'm sure this seems trivial to many people, but it's worth getting
these names right (and doesn't hold anything else up, as they're easy
to change at any time).

If we say "tree" instead of "subtree", does that address that
particular issue, Max?  It's still close enough to the LDAP original
that we'll get the benefit of familiarity.

"Base" vs "zero" is a little tougher.  I personally don't find either
one much more intuitive than the other, and would like not to differ
unnecessarily from *some* standard name -- in this case, the LDAP
name, for consistency with the "onelevel".

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Objection to change in svn_depth_t component naming (r21067)

[David Anderson <da...@calixo.net>]

* Karl Fogel <kf...@google.com> [2006-08-14 17:55:45]:
> Why would we assume that the DAV standard is more "established" than
> the LDAP one?  We happen to be unusually familiar with DAV, but that's
> probably just a local bias...

You said it: Subversion has a local bias to use DAV terminology. If we
start mixing in LDAP terminology, wouldn't this result in more
confusion than anything else?

I also feel that the DAV terms are much more explicit than the LDAP
ones (could be that dav bias again).  Does it not suffice if the
corresponding LDAP terms are documented in the design?

- Dave.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Objection to change in svn_depth_t component naming (r21067)

[Max Bowsher <ma...@ukf.net>]

Karl Fogel wrote:
> Max Bowsher writes:
>> All-in-all, I'd much rather we stuck with the established DAV
>> standard, and ignored the LDAP terminology completely.
> 
> Why would we assume that the DAV standard is more "established" than
> the LDAP one?

Sorry, I was unclear. I didn't mean to compare the two standards. 
Rather, I was referring to the fact that we already had the code using 
the DAV terminology prior to r21067.

> We happen to be unusually familiar with DAV, but that's
> probably just a local bias...

Yes, but these are developer-visible terms, are they not? So, falling in 
line with local bias seems like a *good* thing.

> How about "fulltree", btw?

It doesn't address my complaints about "base" or "exact", whilst 
simultaneously deviating confusingly from the LDAP standard.

Moreover, let me clarify my position: I am opposed to using various 
'names' for certain patterns of depth, when (IMO) depth is a concept 
best described in numerical terms - i.e. from my perspective, *any* set 
of descriptive words is inferior to the DAV-style "zero", "one", "infinity".

Max.

Re: Objection to change in svn_depth_t component naming (r21067)

[Karl Fogel <kf...@google.com>]

Max Bowsher <ma...@ukf.net> writes:
> All-in-all, I'd much rather we stuck with the established DAV
> standard, and ignored the LDAP terminology completely.

Why would we assume that the DAV standard is more "established" than
the LDAP one?  We happen to be unusually familiar with DAV, but that's
probably just a local bias...

How about "fulltree", btw?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Objection to change in svn_depth_t component naming (r21067)

[Max Bowsher <ma...@ukf.net>]

Karl Fogel wrote:
> Max Bowsher <ma...@ukf.net> writes:
>> Sorry, I wasn't following this thread and didn't comment before - I've
>> just noticed that such a change was committed in r21067.
>>
>> I would like to object on the grounds that the zero/one/infinity
>> naming scheme is far more intuitive than the LDAP names.
>>
>> This isn't just a reflexive action the change on the first time
>> encountering the LDAP names - I've run into them before, and found
>> them unnecessarily confusing about the precise concepts they refer to.
>>
>> Whilst "onelevel" is reasonable, neither "base" nor "exact"
>> communicate their concept to me as clearly as "depth zero". Ditto for
>> "subtree", which makes me think of selecting just a particular subtree
>> of nodes 
>> From underneath the object it is being applied to, not entire tree 
>> rooted at the reference object.
> 
> I'm sure this seems trivial to many people, but it's worth getting
> these names right (and doesn't hold anything else up, as they're easy
> to change at any time).
> 
> If we say "tree" instead of "subtree", does that address that
> particular issue, Max?  It's still close enough to the LDAP original
> that we'll get the benefit of familiarity.
> 
> "Base" vs "zero" is a little tougher.  I personally don't find either
> one much more intuitive than the other, and would like not to differ
> unnecessarily from *some* standard name -- in this case, the LDAP
> name, for consistency with the "onelevel".

"tree" is marginally better than "subtree" in that it is no longer 
communicating an implication of sub-selection, but (to me) it is still 
unclear, since there is no explicit indication of how far down the tree 
you wish to travel - "onelevel" is also a tree, after a fashion.

The concepts of "base depth" or "exact depth" have no intuitive meaning 
to me, whilst "zero depth" does.

All-in-all, I'd much rather we stuck with the established DAV standard, 
and ignored the LDAP terminology completely.

Max.