Re: Comments/Questions on draft-ietf-deltav-versioning-06

From: Geoffrey M. Clemm (geoffrey.clemm@rational.com)
Date: Thu, Jul 27 2000

  • Next message: Geoffrey M. Clemm: "Re: Mini scenario"

    Date: Thu, 27 Jul 2000 16:46:11 -0400 (EDT)
    Message-Id: <200007272046.QAA21708@tantalum.atria.com>
    From: "Geoffrey M. Clemm" <geoffrey.clemm@rational.com>
    To: ietf-dav-versioning@w3.org
    Subject: Re: Comments/Questions on draft-ietf-deltav-versioning-06
    
    Hi Gary!
    
    Thanks for the review.  A new perspective is always much appreciated!
    
       From: karasiuk@ca.ibm.com (by way of "Ralph R. Swick" <swick@w3.org>)
    
       3.4.1 - The sentence "This property contains the revision URL for this
       revision" is quite confusing.  It sounds like it is just pointing back to
       itself? Could this please be expanded.  If it is some internal url, then
       perhaps it needs a different name, because in other parts of the draft
       (e.g. 3.3.1) you talk about the "revision URL" and so then it not clear
       what you mean (do you mean this property or this element?)
    
    The term "revision URL" always refers to a URL that identifies a
    particular revision.  This revision URL will always identify this
    revision, unlike a "versioned resource URL", which will display
    the content and properties of whatever revision is indicated by
    its DAV:target property.
    
       3.5.1 - Why is there the indirection for the checked-out property.  Why not
       just <!ELEMENT checked-out (href)>  ?
    
    This just embeds a semantic constraint in the syntax (i.e. that the URL
    must identify a revision), but I'm happy to just have it be an "href"
    if people care.
    
       5.3 - The whole topic of selection is confusing.  What if you don't specify
       a target selector, what do you get?
    
    Good point!  I'll add the following Postcondition:
    
    "If the request-URL identifies a versioned resource, the response
    will contain the content of the target of the versioned resource."
    
    Similarly for PROPFIND.
    
       Which of the other operations update the default target? (checkin?
       checkout?).
    
    Only the ones that explicitly state that they do (i.e. CHECKIN and
    SET-TARGET).  I just checked the definition of DAV:target, and it
    neglects to mention CHECKIN ... I'll fix that!
    
       And if you say the default is the "latest", then what percisely do
       you mean by the latest?  the revision with the highest timestamp?
    
    
    There is no "latest" target.  There is a "DAV:latest-checkin-report"
    which returns the revision(s) with the latest checkin date.
    
       6.2 - It would be useful to add a sentence that explains that check
       out doesn't actually return the contents/body of the checked out
       thing (because most other source control systems do return the body
       on checkout).
    
    Normally, we only specify what a method does, not what it does not do
    (after all, there are an infinite number of things it does not do :-).
    
    In this particular case, I'm not aware of any SCM systems that return
    the body on checkout (they just return a status value, i.e. "success"
    or "failure").  What SCM systems did you have in mind?
    
    Cheers,
    Geoff