Conference vaxaxp::vmsnotes

The HELP for SET PROTECTION/DEVICE states it's for non file-structured devices
which would exclude disks.

(it also states it's been replaced by SET SECURITY/CLASS=DEVICE)

    
    To expand on the comments made in .1, setting the protection on a file 
    structured device is frequently misunderstood. The device protection is
    only relevent when no volume is mounted, and therefore only effects
    operations on the unmounted volume. Normally the only operation we'd
    expect to allow is mounting the volume.
    
    Perhaps you mean to set the volume protection? 
    
    (in releases prior to V6.0, VMS tended to blur this distinction, which
     created problems with our security model)

    I am refering to the device. I am trying to make the device available
    to other users so that they can mount, dismount, and perform other
    such operations without requiring access via the SYSTEM account.
    
    I finally figured out that the command required is of the form...
    $ set security/class=device/protection=(w:rwpl) TARDIS$DKC100:
    
    This required some figuring out since on-line documenation on security/class
    does not list any qualifiers as options and offers not examples.
    
    I'm going to QAR the entry in on-line help for set/protection,
    because although there is a note stating the command is superceded by
    security/class, the format given is in the obsolete syntax:
    
    SET PROTECTION=(ownership[:access][,...])/DEVICE device-name[:]
    
    This is where I strayed from the path: I initially typed in help
    set protection/device and went straight to the format shown.
    
    I am surprised to find that format was not replaced with
    something like: 
    try  $ set security/class=object_class/protection=(...) device_name
    
    If a command is obsoleted, I believe the documentation should
    provide more detail.

    I also believe that the help is lacking in detail as I've already
    stated. 
    
    I am curious why commands are obsoleted and replaced altogheter 
    rather than simply extending the DCL interface.
    
    Anyway, thanks for the help. - Mark
                                    

:    I finally figured out that the command required is of the form...
:    $ set security/class=device/protection=(w:rwpl) TARDIS$DKC100:

   Depending on the intended amount of accessability to general users
   on the system, an ACL might also be of interest.
    
:    I'm going to QAR the entry in on-line help for set/protection,
:    because although there is a note stating the command is superceded by
:    security/class, the format given is in the obsolete syntax:
:    
:    SET PROTECTION=(ownership[:access][,...])/DEVICE device-name[:]

   It's likely best that the whole section -- past the "See SET
   SECURITY" -- be pulled...

:    If a command is obsoleted, I believe the documentation should
:    provide more detail.

   The online HELP is intended as a quick reference -- we tend to
   pull obsolete commands and routines out of the documentation,
   and out of the on-line HELP.  (Amusingly, the Obsolete Features
   Manual is itself considered obsolete.)

:    I also believe that the help is lacking in detail as I've already
:    stated. 

   This is expected -- the BOOKREADER and WWW documentation libraries,
   as well as the hardcopy manuals, contain the full documentation.
    
:    I am curious why commands are obsoleted and replaced altogheter 
:    rather than simply extending the DCL interface.

   SET SECURITY replaced a selection of DCL commands, each with
   unique syntax...  Given your node, there are a number of folks
   in your immediate area that can discuss this issue...

re .3
    
>    I'm going to QAR the entry in on-line help for set/protection,
>    because although there is a note stating the command is superceded by
>    security/class, the format given is in the obsolete syntax:
>    
>    SET PROTECTION=(ownership[:access][,...])/DEVICE device-name[:]
    
    If we are going to document SET PROT at all then it needs to show its
    own format *and* refer to replacement commands.
    
    I would be concerned if a command is in the help but refuses to tell
    the reader what its format is. [If the old command no longer works,
    then this is probably a legitimate QAR in its own right.]
    
    Your QAR should include a comment that SET SEC/CLASS does not list the
    possible valid classes.
    
>    I also believe that the help is lacking in detail as I've already
>    stated. 
    
    I suspect this is for historical reasons. When VMS was first shipping
    the size of disks was such that one would not want to have all the
    documentation online. Today there is no reason not to make the full
    doco available online (eliminating the distinction between the online
    help and the more comprehensive dead tree version) - except perhaps
    that we sold our disk business.
    
>    I am curious why commands are obsoleted and replaced altogheter 
>    rather than simply extending the DCL interface.
    
    Evolution leaves behind mess. The new SET SEC command represents the
    idea that all the different aspects of security (UIC, protection, ACL)
    and all the different types of objects (file, device, ...) are ideally
    controlled in a unified way. One could not extend the previous commands
    to do this (SET FILE, SET PROT, SET PROT/DEVICE, SET ACL) in a clean
    way but of course the old commands remain so as to avoid breaking
    existing procedures.
    
    As .-1 says, if you are on STAR:: then you have ready access to those
    who can answer your every question.

    
:>    I also believe that the help is lacking in detail as I've already
:>    stated. 
    
:    I suspect this is for historical reasons. When VMS was first shipping
:    the size of disks was such that one would not want to have all the
:    documentation online. Today there is no reason not to make the full
:    doco available online (eliminating the distinction between the online
:    help and the more comprehensive dead tree version) - except perhaps
:    that we sold our disk business.

   The existing help "engine" is twenty-year-old technology, and is
   completely unweildy for any decent volume of information.  The
   more one puts into help, the harder it is to find anything...

    re .6
    
    I hope I didn't give the impression that I thought that in making *all*
    the documentation online we should retain the current $ HELP command.
    
    I would prefer an HTML based solution over PDF because the latter is
    proprietary and I happen to find my PDF reader a bit clunky but that's
    just MHO.

:    I would prefer an HTML based solution over PDF because the latter is
:    proprietary and I happen to find my PDF reader a bit clunky but that's
:    just MHO.

   HTML is already available.  If you want to have a mirror site for
   the latest documentation updates -- we have been looking for one on
   the other end of the planet -- I can get you in touch with the folks
   maintaining the continuously-updated HTML documentation version.

   I'd like to see a non-clunky PDF viewer for OpenVMS either from
   DIGITAL or from Adobe -- PDF is a pretty good solution for this
   sort of situation...

>        <<< Note 413.8 by XDELTA::HOFFMAN "Steve, OpenVMS Engineering" >>>

>    I'd like to see a non-clunky PDF viewer for OpenVMS either from
>    DIGITAL or from Adobe -- PDF is a pretty good solution for this
>    sort of situation...

In particular, PDF can do a better job of presenting the layout
intended by the author/designer than HTML-based approaches.

The fact that HTML is ubiquitous does not mean it is perfect.

re .8
    
>   I'd like to see a non-clunky PDF viewer for OpenVMS either from
>   DIGITAL or from Adobe
    
    The PDF viewer that I was using is Acrobat for Mac. I guess it offers
    cross-platform clunkiness.
    
> If you want to have a mirror site
    
    I would be interested. The machine that I would want to use for this
    does not currently have an HTTP server installed. Is there one
    available free (internal use only) for Alpha VMS V6.2? I would want to
    get that sorted out before putting up my hand.
    
re .9
    
    I guess this is a philosophical difference in HO. I would regard HTML's
    lack of integrity to the minutiae of presentation a strength, not a
    weakness.
    
    I would rate as the most important aspect for online (or offline for
    that matter) information, whether one can find what one is looking for.
    
    While style (e.g. readability) is important, we already know that for
    the VMS doco it is reasonable or better and that would be preserved by
    most solutions. I believe that HTML provides sufficient control over
    layout.
    
    Accuracy of information is also something that I would rate above
    exactitude of presentation.
    
    I can imagine that the priorities might be different if we were
    delivering marketting brochures rather than technical documentation.

    
    I agree that ubiquity is no guarantee of quality.