Re: Requirements in release

[Daniel Fetchinson <fetchinson googlemail com>]

>> >> >> Why can't the packaged archive contain a reasonable configure
>> >> >> script to begin with?
>> >> >> [...]
>> >> >> In any case the user experience could be increased greatly if
>> >> >> the installation of sawfish would not involve all sorts of
>> >> >> trickery, but the simple
>> >> >> ./configure
>> >> >> make
>> >> >> make install
>> >> >>
>> >> >> cycle on each of the dependencies (librep, rep-gtk, sawfish).
>> >> >
>> >> > ... well, normally it is shipped, somehow I forgot it for 0.18.6
>> >> > (...)
>> >>
>> >> Daniel Fetchinson is right.
>> >>
>> >
>> > yep.
>> >
>> >> Chris, I'm sorry that we may rely too much on you, but the release
>> >> is
>> >
>> > :/ sometimes I don't know if that is because you trust me that
>> > much, or if you are moving away from sawfish, just like Sven (and
>> > others) did.
>> >
>> > (Ehh, you is plural here, not you as in Teika)
>> >
>> >> important, especially "big ones" like sawfish-1.5. I wonder if you
>> >> could prepare, say, a checklist of release procedure. (Of course it
>> >> can contain a line or two of "Press Teika to do this tiresome
>> >> job." I can't refuse. Ouch ;) Because there's an interval between
>> >> big version changes, it's easy to forget this and that and etc.,
>> >> etc. It can also help at the maintainer-ship transfer.
>> >>
>> >
>> > Well, this checklist does (imaginarly) exist, but is not failsafe,
>> > I'm just happy that it "just" happened to rep-gtk 0.18.6, as there
>> > are only changes to the specfile (nothing else) and 0.18.4 is
>> > enough for sawfish. It would have hurt much more on librep or
>> > sawfish. But of course you're right, that this should not happen
>> > again.
>> >
>> > ...
>> >
>> >>
>> >> One more thing I'd like to point out is that the news items are
>> >> terse. Just a delay of two weeks or so can make it far better.
>> >> Clearly what lacks though there should be is the tab usage.  In
>> >> fact, I was upset by the 1.5 release, and I quickly fixed a small
>> >> manual on wiki: http://sawfish.wikia.com/wiki/Tab
>> >> At least, there should be the pointer in news.texi to it. Please
>> >> put it on 1.5.1 / 1.6.
>> >> (Yes, focus-revert is important, too, but to make it short, I skip
>> >> it now.)
>> >> Rewinding time a week, I was about to write the last cleanup of
>> >> NEWS for 1.5.
>> >>
>> >
>> > I'm sorry Teika, I wanted to do the release finally, as it already
>> > delayed 4/5 weeks. But well, any improvement on this is -of course-
>> > welcome, and for the next time: 1.6.0 is definitively to be
>> > released on the 22nd of December, so feature freeze is end of
>> > November and we can then better handle stuff like that. For 1.5.x
>> > which is "only" a bugfix series this may not be that important
>> > (except that stuff from 1.5.0 that's to be improved).
>> >
>> >> For ordinary softwares, I expect things go out-of-box, or at least
>> >> there accompanies a decent manual. If not, I curse, "d***, [...]".
>> >>
>> >> Because I like sawfish, this is not the case for me on sawfish,
>> >> but PLEASE LISTEN, many users swear, and for them sawfish is not
>> >> reputable.
>> >>
>> >
>> > You're right.
>> >
>> >> Popularity is the power. Some reports, what's good and bad (or
>> >> bugs). They help. Some subscribe ML. A few contribute in ML. A few
>> >> of few develops.
>> >>
>> >> Anyway, thanks a lot for all. I was happy when I knew Sawfish was
>> >> resurrected. Now we see a boom come. Another will come.
>> >>
>> >
>> > A boom? You mean much more Sawfish users? Sorry, but if you're
>> > thinking this way, I think you're wrong:
>> >
>> > GNOME 3 will not allow to use sawfish anymore! The GNOME Shell will
>> > only work with this ultra-ugly MetaCity shit. So only KDE will be a
>> > Desktop usable with sawfish. So we either wait for Timo to get KDE4
>> > running, check if we need to do some improvements for better KDE4
>> > Support and go trying to get some KDE-Users to use sawfish,
>> > or build our own desktop, nothing like GNOME/KDE, but like LXDE.
>> > Standalone Windowmanagers are not suitable for most users theese
>> > days.
>>
>> Well, I actually have been using sawfish standalone without gnome or
>> kde since it appeared in redhat. And I absolutely would like to keep
>> using it as such in the future, regardless of what gnome and kde comes
>> up with and I also don't care about so called minimalist desktops like
>> lxde. I think sawfish by itself is just great and if it ain't broken
>> don't fix it. The only extra thing I use is a trayer application,
>> either one by the name 'trayer' or 'stalonetray'.
>>
>> > Oh well, and in the last half year there have more unsubscriptions
>> > than subscriptions, but more users are monitoring
>> > librep/rep-gtk/sawfish on sf.net, the IRC is pratically dead (I
>> > only have had 2 (in words: two) discussions there the last half
>> > year), so I would say: The userbase did not change that much.
>> >
>> > Common things noobs think about sawfish (not a joke):
>> >
>> > - Too many functions, which no one needs (the argument that you just
>> >   don't know them and therefore can't say if you would need them is
>> > not of interest)
>> > - Bad Usability (well this is that sawfish-ui is shit therefore
>> >   sawfish is also shit thingy)
>> > - It's not part of most Distros - can't be good
>> >
>> > Of course this is only a group of users and not all.
>> >
>> > This may sound pesimistic, but: I absolutely don't care how many
>> > users are using Sawfish, for me it's the Ultimate Windowmanager,
>> > regardless of anyone else, so you don't have to fear to get rid of
>> > me :p
>>
>> I think what would attract the right kinds of users is a clear message
>> about the intended audience. It should be clear from the project page
>> who are expected to use sawfish and for whom probably it's not a good
>> choice. If this policy is clear then the effort can go into convincing
>> the right kinds of users and no effort will be wasted on preaching to
>> people who will not care anyway. For example it might be the case that
>> gnome 3 is not compatible with sawfish anymore, but is this a problem?
>> I don't think so. I don't know how many gnome users use sawfish but
>> I'd think much more are using sawfish standalone.
>>
>
> Well Sawfish used to be the GNOME Window Manager, so there's still a
> bunch of them using GNOME.

I estimate that users who use both gnome and sawfish are very few
these days although I can't be sure of course. What I am sure though
is that the number of such users will very soon go to next to zero. In
other words any effort and energy that goes into making sawfish work
well with gnome is pure waste.

> But well who's the "intended audience" of
> sawfish? I would say: Advanced Users who want to have full control over
> their WindowManager. Mainly, but not limited to theese.

Yes, I agree. I would also add that your definition of "intended
sawfish audience" excludes gnome and kde users, or if it doesn't
exclude them yet, they will soon do. This is simply because gnome and
kde users do not want to have full control over their window manager.
It's nothing bad, they can be programming wizards and advanced users,
but as far as their Desktop Environment goes, they want gnome or kde
to take care of everything.

Because of this I think not only gnome, but kde integration of sawfish
is a waste of time. Of course if somebody wants to use kde with
sawfish and as a result dedicates time to development, that's fine.
He/she can do whatever he/she wants with his/her free time. But it's
important to realize that this will not bring in new users or new
developers.

>> An example: I'm pretty sure much more people use vim in text mode than
>> in graphical mode together with the mouse. Vim has some graphical
>> features but the main development effort and marketing effort goes
>> into the text mode version, which is considered to be the 'real vim'.
>> And vim is really attractive and great and popular among these kinds
>> of users.
>>
>
> I'm also a Vim User. Textmode of course :)
>
>> Similarly, the main effort I think should go into making sawfish as
>> good as possible for standalone mode because these users are the
>> natural audience who are really committed to sawfish.
>>
>
> Of course, this "Desktop Environment" thing I spoke about is exactly
> this, while DE does mean: panel (with menu/notificationarea the later
> is important for me/pager/windowlist), background setter, support for
> desktop icons, those 3 without relying on 3rdparties. But well, of
> course it's already all possible with 3rdparties. Perhaps it's just my
> point of view.

Exactly these things are the ones I absolutely don't care about
(panel, menu, notification area, pager, background setter, desktop
icons, etc) and don't use them at all. If I would care about these I
would use gnome, kde or a minimalist desktop environment. The reason
is that these software projects solve these problems very well, there
is no need to reinvent the wheel.

But sawfish is unique in the sense that as a standalone window manager
it's very flexible and great to use. This aspect should be reinforced,
features stabilized, debugged, etc. All the rest should be I think
left to software projects whose main focus is that. In those area
sawfish can not compete. So it should not go into areas like that. It
should stay where it is competitive, because there is no other great
window manager that is standalone, fast and very flexible.

>> Looking forward to a long lifespan for sawfish!
>>
>
> You're looking into a shiny future :)

Cheers,
Daniel


-- 
Psss, psss, put it down! - http://www.cafepress.com/putitdown