Woof road map?

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Woof road map?

mavrothal
There is already a thread for the immediate (slacko) woof roadmap ( http://woof-ce.26403.n7.nabble.com/Slacko-and-woof-road-map-for-immediate-future-td243.html ) but I was wondering of a readmap more on core puppy elements.
The way I see them starting from power-on, these are boot sequence (UEFI), hardware compatibility (pup_event/udev/systemd), X/GUI initialization (xorgwizard), core desktop (gtk) and window manager, connectivity, package manager/binary compatibility, multi-user support.
There are also some ARCH issues ie ARM support.

I believe that the boot sequence (initrd/init) and the range of install options that it provides it is a puppy characteristic and it only may need bug fixes. Regarding UEFI support I think we'll be getting there soon.

With all major distros yielding to RHEL/systemd puppy may soon have a problem specially when using packages from other distros. Hopefully eudev will keep developing to allow more "traditional" kernel-iron communication but the dependencies issue must be addressed somehow. Any thoughts on this?
Regardless of udev/systemd though I think that the pup_event infrastructure is becoming obsolete and I think that desktop drive icons is the last thing hooding it in place. Pup_volume_monitor is almost there and I think it could become default in puppies.
 
On X/GUI bring up, I believe Xorg/xrandr are doing a pretty good job and xorgwizard is becoming less relevant and it could be removed.

On the desktop front. ROX/JWM is another puppy-staple and I believe a good one. Would be nice if hard-coding to ROX as a desktop manager could be minimize, but given ROX-apps ROX is destined to stay with puppy I think. What we really need is someone to pickup ROX development again. I'm wondering what techno is doing these days…
Gtk3 is another UI issue. I believe it sucks and puppy may want to give a more serious look to Qt for the future. But for now I think Gtk2/gtkdialog should remain puppy's core UI infrastructure till it drops dead ie till major apps like Browsers, do not run on it.

Connectivity wise I think puppy is OK with the current apps, and with the aforementioned possible removal of pup_event and changes in firmware location and initialization will probably be OK. One change may be the drop of analogue modems and better support for bluetooth. I really do not know how many puppy users will be using a modem a year or 2 from now, but I can imagine that most will have some bluetooth device.
 
Although, PPM and the package system in puppy has received considerable criticism through the years, I believe is the only one that can handle the wide range of packages that puppy uses and it has been improving in handling dependencies. I think it will mostly need bug-fixes as they appear rather than revamping.
Binary compatibility wise, it is clear that puppy is doing better with ubuntu/debian and slackware. I'm not sure if puppy was ever intended or wants to be a breeding edge (:-o) distro but if it does, Fedora/rpm and Arch support must be improved. I personally do not find it important. Quite the opposite. I think that adding another distro or 2 does not add anything to the core puppy values and it actually regenerates the GNU/Linux dispersion within a single distro. But that's me…

A final point is multi-user support/running as root. Puppy s designed as single-user system that can run from CDs and USBs, and thus root is both mandatory and default. I am personally not sure if puppy should become a genuine multi user system in the sense that the required changes may hamper some of its characteristics as speed, efficiency and simplicity.

However, puppy is a do-ocracy.  
All of the above are irrelevant unless someone starts working on them.
I put them here not so much to initiate (yet another) discussion but as a trigger for developers to prove them right or wrong with their code ;-)
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Woof road map?

q5sys
I've been out of the loop with Puppy stuff for a while now due to real
life needing to take focus. I'm sure everyone has been there and
understands.  My contributions over the last few months have been very
minimal due to work and family life, but there were a few things here
have come up in thought before.

> Gtk3 is another UI issue. I believe it sucks and puppy may want to give
> a more serious look to Qt for the future. But for now I think
> Gtk2/gtkdialog should remain puppy's core UI infrastructure till it
> drops dead ie till major apps like Browsers, do not run on it.

I cant recall exactly when and with who, but I believe at least the
comment about a Qt based Puppy has been made before.  I know it wasnt
here, might have been on Murga or through an email with someone. That
would be a very interesting development, but sadly I think it's beyond
the scope of my abilities. If I get free time I will attempt to tinker
in that area, but I can't promise much.  I do think though that it could
eventually lead to us being able to do some work towards fighting the
'Puppy is Ugly' comments I hear almost every day.
Not that default UI appearance is really a top thing on our priority
list, but its still a thorn in the side when trying to promote Puppy and
that's usually alwyas the top reply.
Anyway moving along.

> One change may be the drop of analogue modems and better support for
> bluetooth. I really do not know how many puppy users will be using a
> modem a year or 2 from now, but I can imagine that most will have some
> bluetooth device.
This one is a tough one for me, because I'm always a big proponent of
legacy support, but even I have to admit that the volume of users who
would actually use dialup has to be next to none.

I agree that Bluetooth support is probably of more use to most people.  
However how much would removing our legacy dialup stuff do for us?  
Would we really gain that much, or would it simply just clean up the
networking menu entries?

> Although, PPM and the package system in puppy has received considerable
> criticism through the years, I believe is the only one that can handle
> the wide range of packages that puppy uses and it has been improving in
> handling dependencies. I think it will mostly need bug-fixes as they
> appear rather than revamping.

As a tool itself, I've never had a major problem with PPM.  I've never
had much of an issue with it and I can agree that it's gotten alot
better with dependencies over the past 5 years.  I think the biggest
improvement we could do really has less to do with the PPM and more to
do with Package Repos themselves.  I've wanted to build a more modern
repo infrastructure for Puppy for a while, but it seems that most dont
seem to mind the complete scattered clusterf*ck we currently have.  (I'm
sure some think everything is just fine and dandy).
I'm still on board with working to make this happen, but there's no
point in me putting out the money and coding effort for a central site,
CDN, etc; if no one else would care to use it.



> Binary compatibility wise, it is clear that puppy is doing better with
> ubuntu/debian and slackware. I'm not sure if puppy was ever intended or
> wants to be a breeding edge (:-o) distro but if it does, Fedora/rpm and
> Arch support must be improved. I personally do not find it important.
> Quite the opposite. I think that adding another distro or 2 does not
> add anything to the core puppy values and it actually regenerates the
> GNU/Linux dispersion within a single distro. But that's me…

I agree that puppy is doing great with Slackware as its base.  I think
debian is a better option going forward than ubuntu, but with debian now
going systemd... we may be facing issues there within the next 5 to 10
years.  Although I'm sure debian, being debian, will not be hard coding
vanilla packages for systemd.  I know there arent many at this point,
but Im sure as more time goes by, more projects will follow the lead of
GNOME and just assume its what everyone will be using.
While I like various aspects of systemd, I still prefer legacy sysvinit.
  Which brings me to the next point on this.  Arch support.  I'll openly
admit that I am a huge fan of Arch, and I do believe that the mentality
and process behind arch fits well with puppy.  Packages are as upstream
as possible and its amazingly pluggable.  With the package availability
in arch it would pretty much end the 'I cant find this app for puppy',
and remove 90% of the dep issues when trying to install something after
the fact.
But I acknowledge the amount of work that would go into making a pure
arch puppy.  I also can admit that the arch community having the
mentality that it does, probably wouldn't be too accepting of it.  
Without bringing up past drama and issues, I'm sure everyone here knows
how brushes with loyal arch believers and their work with puppy has been
rough for us.
And as much as I'd love a pure arch puppy, I'd rather forgo that and
keep our community happy and working together.

> A final point is multi-user support/running as root. Puppy s designed
> as single-user system that can run from CDs and USBs, and thus root is
> both mandatory and default. I am personally not sure if puppy should
> become a genuine multi user system in the sense that the required
> changes may hamper some of its characteristics as speed, efficiency and
> simplicity.

My question here is what do you mean by multi-user support?  I know it
may sound dumb, but I've heard alot of people say 'multi-user/root' but
really mean 'non-root/root'.  Are you talking about the former or the
latter?  Due to puppies design and for that matter, core goal... I dont
see how 'multi-user' is of much benefit.  I've never cared about the
multi-user support thing as much as some others, I just tell someone to
reboot and choose a different save file for example.
The non-root/root issue is one that i have no problem with if we change,
and I do have to say I like the implementation that JamesBond and Kirk
have worked into Fatdog64.

> However, puppy is a do-ocracy.
> All of the above are irrelevant unless someone starts working on them.
> I put them here not so much to initiate (yet another) discussion but as
> a trigger for developers to prove them right or wrong with their code
> ;-)

Understood and agreed.  I hope I can rein my life back under control so
that I can get involved a bit more... sometimes life just gets in the
way of what we really want to do. haha

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Woof road map?

scsijon
On 04/05/2014 08:49 PM, q5sys [via woof-CE] wrote:

>
>
> I've been out of the loop with Puppy stuff for a while now due to real
> life needing to take focus. I'm sure everyone has been there and
> understands.  My contributions over the last few months have been very
> minimal due to work and family life, but there were a few things here
> have come up in thought before.
>
>> Gtk3 is another UI issue. I believe it sucks and puppy may want to give
>> a more serious look to Qt for the future. But for now I think
>> Gtk2/gtkdialog should remain puppy's core UI infrastructure till it
>> drops dead ie till major apps like Browsers, do not run on it.
>
> I cant recall exactly when and with who, but I believe at least the
> comment about a Qt based Puppy has been made before.  I know it wasnt
> here, might have been on Murga or through an email with someone. That
> would be a very interesting development, but sadly I think it's beyond
> the scope of my abilities. If I get free time I will attempt to tinker
> in that area, but I can't promise much.  I do think though that it could
> eventually lead to us being able to do some work towards fighting the
> 'Puppy is Ugly' comments I hear almost every day.
> Not that default UI appearance is really a top thing on our priority
> list, but its still a thorn in the side when trying to promote Puppy and
> that's usually alwyas the top reply.
> Anyway moving along.

I think that was a murga thread and me some time ago, I was planning a
qt2 ?or qt3 based puppy based on my myz puppy (minimal racy), but had to
stop as there was just too many problems with implementing it due to
puppy's ways of working, package versions it used and the overall
packageset available at that time. There was an alpha I released back
then, but that was the end of it. However with the latest puppy's it can
be implemented as I know someone had a qt package thread going on murga
a few months ago when I scanned, and it seems that the qt people have
got their act together with backward compatability conversioning rather
than you having to manually 'fix' your sources for each sub-version update.

>
>> One change may be the drop of analogue modems and better support for
>> bluetooth. I really do not know how many puppy users will be using a
>> modem a year or 2 from now, but I can imagine that most will have some
>> bluetooth device.
> This one is a tough one for me, because I'm always a big proponent of
> legacy support, but even I have to admit that the volume of users who
> would actually use dialup has to be next to none.
>
> I agree that Bluetooth support is probably of more use to most people.
> However how much would removing our legacy dialup stuff do for us?
> Would we really gain that much, or would it simply just clean up the
> networking menu entries?

And yet there are pleanty of ISP's that still have dialup available, and
out here in Victorias Mallee it's still available at quite a cheap cost.
IF the NBN ever get their act together it will be redundant, but
pollies.....

jon
Loading...