Re: Simpler name for --no-clobber option

[Tomas Zubiri]

*Is GNU an api or a user interface?*
The reason for not including this change is reasonable and expected
resistance: " While the new name may be clearer
scripts using it would not be compat with older cp implementations.
Hence the incompatability introduced would not be worth it."

I've never had to worry about the version of cp, so I appreciate the
constant nature of gnu. I also never used any cp option in any script, and
if I would it would be in a context where the operating system is
versioned, and containered or virtualized, so I would not be impacted
anyway.

What I think is going on here is that GNU originally had the goal of being
a tool for end users, but this was incompatible with its de facto usecase
as a software foundation, where the demographic of its interface is people
who build software for end users. For this usecase, its constant,
versionless, dependable nature is actually a positive attribute.

I think one of the advantages Windows had over Linux/gnu is that their user
interfaces, by design, could not be used as the foundation of another
system. There is a distinction between an api and a user interface, so the
user interface could change without impacting programmatic consumers. Is cp
part of an api or a user interface? It seems to be both.

However the api nature of GNU is stronger because it's made by engineers,
so naturally they better serve engineer users, we can clearly see this
phenomenon in action in the wording of clobber instead of replace, it seems
clobber "in the context of data manipulation means destroying data by
overwriting it". Something I learned from the following stackexchange
question
https://unix.stackexchange.com/questions/572294/why-is-cps-option-not-to-overwrite-files-called-no-clobber?noredirect=1#comment1065230_572294
p.s: along with the fact that mv also has a no clobber option. Another
instance of this end-user engineer-user dichotomy is the fact that
no-clobber was implemented by a redhat engineer (who also responded to this
thread, kudos for the constancy in support to both redhat and the
engineer!).

I'd be tempted to redirect my efforts towards one of the GUI projects that
build upon coreutils. But I like clis even as a user. It's just that, if
they are as conservative and immune to change like this, they will never
improve.

Is GNU content with serving other engineers and given up on the goal of
serving end users? I, like I assume the founders did, have dreams where
users can type something like 'copy from myfolder to usb', and the system
responds with 'do you mean the usb folder or your usb drive?'. But we can't
reach that dream if it can't support something as basic as a synonym. Am I
in the right place?

Kind regards,
Tomas Zubiri


Tomas Zubiri
Software Developer
Cel: +54 911 5891 8239



El mié., 11 de mar. de 2020 a la(s) 16:27, Leslie S Satenstein (
address@hidden) escribió:

> Did you check the dictionary meaning of clobber?   I am happy to use it,
> but clobber means to abuse physically.
>
> Merriam Webster dictionary
>
> Definition of *clobber* (Entry 2 of 2)
>
> transitive verb <https://www.merriam-webster.com/dictionary/transitive>
> 1 *: *to pound mercilessly also *: *to hit with force clobber a home run
> 2a *: *to defeat overwhelmingly
> b *: *to have a strongly negative impact on businesses clobbered by the
> recession
> c *: *to criticize harshly
>
> clobber does not appear to non-abstractly match overwrite.
> Ignore my response.  I go with the flow
>
> Regards
>
>
> * Leslie*
> *Leslie Satenstein*
> *Montréal Québec, Canada*
>
>
>
> On Wednesday, March 11, 2020, 11:39:08 a.m. GMT-4, Kamil Dudka <
> address@hidden> wrote:
>
>
> On Wednesday, March 11, 2020 3:37:06 PM CET Leslie S Satenstein via GNU
> coreutils General Discussion wrote:
> > Just saw this message.    for --no-clobber,  would        --no-replace
> > suffice?  the latter could also be shortened to --nr.
>
> I am afraid that you are 11 years late to this discussion...
>
> --no-replace was proposed originally:
>
>     https://lists.gnu.org/archive/html/bug-coreutils/2008-12/msg00157.html
>
> --no-clobber and --no-overwrite were suggested by the reviewers:
>
>     https://lists.gnu.org/archive/html/bug-coreutils/2009-01/msg00053.html
>
> --no-clobber won in the end:
>
>     https://lists.gnu.org/archive/html/bug-coreutils/2009-01/msg00063.html
>
> I do not think that the circumstances have changed so significantly since
> then to warrant a new option supported in parallel to the existing one.
>
> Kamil
>
>
> > Regards
> >  Leslie
> >  Leslie Satenstein
> > Montréal Québec, Canada
> >
> >
> >
> >    On Wednesday, March 11, 2020, 10:32:19 a.m. GMT-4, Pádraig Brady
> > <address@hidden> wrote:
> >  On 11/03/2020 10:08, Tomas Zubiri wrote:
> > > Hello,
> > > I'm attaching a git patch for a small, easy improvement that would
> > > make cp easier and more pleasant to use and learn.
> > >
> > >  cp overwrites files by default, this can be disabled by using the
> > >
> > > command line option --no-clobber, the proposal is merely to introduce
> > > an alias called --no-overwrites.
> > >
> > > Please let me know if somebody would be willing to merge this in
> > > principle so we can move address fine details like testing and the
> > > technical approach used.
> >
> > > For convenience, I'm pasting the patch below as well:
> > While the new name may be clearer
> > scripts using it would not be compat with older cp implementations.
> > Hence the incompatability introduced would not be worth it.
> >
> > thanks,
> > Pádraig
>
>