Puppy's future

What features/apps/bugfixes needed in a future Puppy
Post Reply
Message
Author
elroy

Puppy's future

#1 Post by elroy »

First off, I apologize if this is in the wrong category. If so, sorry Flash.

Just a few observations that I've made, now that BarryK has announced his 2nd retirement...

BarryK appears to be moving on to ARM development, smartphones and tablets. That may be well and good, but I own a tablet (which is based on Android, and an Android based smartphone, too), and I'll be damned if I wipe out the current operating system that it provides, along with a plethora of free-apps, for a niche distro. I guess I can see using Puppy on Raspberry, but again, it is a 'niche' distro...not too much there in terms of a user base. Raspberry is cool, cliche, but not so much mainstream. And I'm willing to bet that it never will be mainstream.

I for one think that Gnu/Linux will survive for another decade or so, desktop or otherwise. Businesses use it, so do developers (even ARM app-developers need desktops/laptops), not to mention those that run servers. And let's not forget the gamers. No, I don't see Gnu/Linux drying up and disappearing in the near future. Will it loose users? Most definitely. Casual users are moving towards tablets. Will it disappear entirely? No, I don't think that will be the case. Even though I have a tablet, and enjoy it immensely, I still find myself using my desktop/laptop at least (if not more than) 80% of the time. There's just so much you can do with a tablet.

And as far as Puppy goes, I'd like to see Puppy move towards a singular distribution (in terms of its base) in order to focus on its strengths. I realize that traditionally Puppy is geared towards older hardware; this is commendable, and should certainly be followed up on. But Puppy should also gear itself towards modern hardware whenever possible.

For my taste it seems that Puppy has become too diverse in terms of its development base. Why do I say this? Well, if you've ever tried to develop on woof2, you'd realize that it doesn't work as advertised. While it may work well with Precise or Slacko, it fails miserably with Wary/Racy, Debian and Arch builds. Arch fails to boot into a desktop (and has a huge ISO size), and Wary/Racy/Debian fail to download much or their required apps, and not just a few...many. It appears that it would be best to concentrate on one thing, which is traditionally the Linux way, and then to do it well. It's all well and good to provide support for Precise, but if BarryK has left the building, how long will that be a viable base for developers using woof2? Same with Slackware? Arch would be a good candidate being that it would use the latest packages because it is a rolling release and woof2 tries to pull in the latest packages, but if you can't even boot into a desktop, and the ISO is huge, what good is it? Wouldn't it be better to base it on a singular distro, say T2 or LFS, in order to make use a dedicated repository? In terms of hardware support, at some point don't you need to say "Ok, here's where the line is going to be drawn...we're going to support computers aged X and newer?". You can't be everything to everyone.

You know, on the one hand I applaud someone like simargl for developing something like his version/vision of ArchPup, which isn't a Puppy at all. But...on the other hand, I perceive him to be an a-hole; I see his attitude to be counter-productive in terms of advancing the Puppy-brand. And you may perceive my attitude to be the same (to each their own). It may be that in the (near) future this may be the future of Puppy development (not the paranoid "he's stealing from me" thing, but more so the totally custom builds that move things away from Puppy).

Well, these are my personal Puppy Linux rants. And my woof2 rants, too :) Maybe you agree, maybe you don't, but at least I've managed to get it off of my chest.

User avatar
darkcity
Posts: 2534
Joined: Sun 23 May 2010, 19:16
Location: near here
Contact:

#2 Post by darkcity »

Maybe Woof could be developed to create better Puppy 'spins'. Maybe it could be combined with adding choices for architectures and packages. I like what porteus has done. allowing you to select options on the the download page-
http://build.porteus.org/

Regarding Woof and Arch people may find this post by Barry informative-
http://bkhome.org/blog2/?viewDetailed=00188
I have fixed Woof to build from Arch packages, however I personally won't take it any further.
...
Firstly, size. My x86 Arch pup build is enormous. This is because Arch developers have given no concern for size in some packages, choosing to link some libraries statically where they need not have been.

Secondly, I am not sold on the rolling release model. For us Linux geeks it is OK, but I don't think it is good for the "great unwashed" (see Sages comment in the previous blog post).
---

There is a wiki page regard Puppy Development but its over a year out of date-
http://puppylinux.org/wikka/PuppyVersionDevelopment
8)

User avatar
Terryphi
Posts: 761
Joined: Wed 02 Jul 2008, 09:32
Location: West Wales, Britain.

#3 Post by Terryphi »

BarryK was willing to spend endless time tweaking Puppy for the enjoyment of his testers. No one else is likely to have the time or inclination to do the same. Frankly, that is not a bad thing. Puppy simply needs a stable core which is supported long term which others can build their variants from if they choose to do so. Is there any agreement on what that core should be?

On my hardware nothing works as well as the T2 builds of Racy/Wary. Does anyone have the ability and inclination to update some of the libraries required by Firefox and Chromium? That is all Racy and Wary need.
[b]Classic Opera 12.16 browser SFS package[/b] for Precise, Slacko, Racy, Wary, Lucid, etc available[url=http://terryphillips.org.uk/operasfs.htm]here[/url] :)

User avatar
Iguleder
Posts: 2026
Joined: Tue 11 Aug 2009, 09:36
Location: Israel, somewhere in the beautiful desert
Contact:

#4 Post by Iguleder »

I agree with elroy.
[url=http://dimakrasner.com/]My homepage[/url]
[url=https://github.com/dimkr]My GitHub profile[/url]

User avatar
6502coder
Posts: 677
Joined: Mon 23 Mar 2009, 18:07
Location: Western United States

#5 Post by 6502coder »

What is the key feature of Puppy Linux that has differentiated it from other Linux distros? To me, it is great support for older hardware. Extending the life of older hardware is what classic Puppy Linux was all about. From that "mission statement" everything else followed: the small ISO size, the focus on lightweight applications with modest hardware requirements, and so on.

Once you throw out support for older hardware as THE key requirement, IMO you no longer have a Puppy Linux. It may be BASED ON Puppy Linux, but it's no longer a Puppy. Once you decide that support for older hardware doesn't matter, then it no longer makes sense to fuss over ISO size and all the rest. Bundle Libre Office as the standard office suite? Sure, why not? Bundle Chrome and Firefox and Opera? Heck, yes. VLC as the standard media player? Of course. GIMP too. Dialup modem support? Why bother? And while we're at it, might as well ditch the whole "devx" SFS idea and just bundle all of that in as well--what's a couple of hundred Megs more?

Don't misunderstand me--I'm not being sarcastic. The need to support older hardware is a severe drag on any attempt to keep pace with the other major distros. So cut loose that anchor. Do it right and you'll end up with a nice, modern Linux distro, full of powerful apps and plenty of bells and whistles and eye-candy. It just won't be a Puppy, not in my book.

That's why Woof--albeit a clever idea and an amazing achievement--always struck me as an exercise in futility. The purpose of Woof was to get access to the rich depositories of major distros like Ubuntu and Slackware. But those depositories would inevitably and increasingly be filled with software that would run poorly if at all on the older hardware that classic Puppy Linux was designed for.

Some have suggested that the definition of "older hardware" should simply be adjusted upward, so that Puppy targets hardware that is (say) ten-years old, at any given time. But most hardware from 2003 is capable of running a full-blown Linux distro anyway. Extending the useful life of that hardware is not such an issue, as is shown by the fact that about a third of all Windows users are still on XP. I sympathize with those who adamantly oppose software bloat on principle--I'm one of them--but let's face it, the average computer user doesn't care, nor should they.

Let's have a clean break. Let Puppy Linux remain focussed on hardware from the Win98 to mid-Win-XP era. Let those who want to move forward set higher, modern minimum hardware requirements so that they aren't handcuffed by the past. But let's not pretend that these modernized derivatives are still going to be Puppies. They have a different charter and should have a different name to match.

User avatar
greengeek
Posts: 5789
Joined: Tue 20 Jul 2010, 09:34
Location: Republic of Novo Zelande

#6 Post by greengeek »

What I want from puppy is:

1) A choice of kernels (to try to match the needs of each of my different machines without being forced to a newer / fatter kernel if I don't want it)
2) Access to programs that run quickly and avoid bloat. (because i feel that bloated programmes increase the risk of backdoor code / phone home code etc etc)
3) Speed
4) Speed
5) Speed

raffy
Posts: 4798
Joined: Wed 25 May 2005, 12:20
Location: Manila

new machines and puppy

#7 Post by raffy »

I've had no problems choosing Puppy to run new machines. The reason is that I define my "new" machine to be low-power and light, the likes of PCs with AMD APUs and Intel Atoms in them. These have always been dismissed by the big boys as "entry-level" hardware, meaning, that only the less demanding (or less able to pay) users would prefer them.

Remember that MS even attempted to douse the enthusiasm for low-power "netbooks" (Atom-based mostly) by limiting its screen size and RAM? That's like throwing the turtle to the river - only this time, the turtle is a Puppy. Umm, no, it's really the rabbit, the fast one. :lol:

So Puppy moves on with the times of changing hardware, and its niche remains to be the people who prefer to use "underpowered" computers. When the eeepc was introduced, Puppy literally was jumping with joy, while the other big OSes were choking.

Now, perhaps we can amend the description of Puppy's target users as those who prefer to use low-power computers ("underpowered", as the big monopolist would put it). That is, in addition to those who keep using old kit.

How do we know that there is such a user base for new but underpowered computers? We look right here. For example, jamesbond is already at Alpha3 of FatdogARM.

Shall this trend of using ARM processors as PC continue? I guess so. Shall I prefer to run Puppy on those machines? Yes, of course!
Puppy user since Oct 2004. Want FreeOffice? [url=http://puppylinux.info/topic/freeoffice-2012-sfs]Get the sfs (English only)[/url].

gcmartin

#8 Post by gcmartin »

Some ideas:

The ISO size decision was made NOT as a RAM decision, but as a download decision. RAM to run PUPs have/has seemingly been available since I joined Puppy Linux. And the PCs many/most of the membership community have and are using work.

RAM size is one of PUPs advantages as it will work with as little or as much RAM as is available on the PCs

Processors and motherboards continue to advance as these provide good and useful speeds as the ever increasing applications exploit hardware for human use and advantage.

There is a PUP, currently, in the kennel for each and every PC that I am aware as having been built by vendor and user. This community demonstrates the widest variety of PC systems; from small to very large in scope.

The concern that we can address and deliver to current and new users is one where this community's developers started several years ago when they began to document the intended PCs and PC needs for the distros they produce for our use. This is a very good policy that they have adopted as it makes easy for any user to see what the developer intends, tested, and visualized for their distros.

Good ideas are being brought to the table in this thread's discussion.

Since time marches on, I think one way of looking at what the OP presents is that it could be time we, as a community, recognizing the change occurring, come up with some new ideas which makes attractive for ANY user to want to use Puppy Linux, regardless of what PC he may have. For example; one statement that could be present somewhere would list distros and some target PCs it is intended. If such a list is present, a quick scan would provide necessary guidance for user selection rather easy. Another example might change the mission statement to embrace the advancements in multimedia that has occurred.

I sure there are other avenues which could be pursued, also, to make Puppy Linux attractive such that it will continue to attract new users with even fresher ideas than we now possess.

For Puppy to flourish, we, like the PC hardware and network services, must also advance, one would think.

But, ...

elroy

#9 Post by elroy »

Yes, some interesting input indeed! I particularly like the fact that we all have an opinion. Diversity is a good thing.

Well, to start with, concerning the belief that Puppy is intended only to be used with older hardware, and the notion that older hardware support is what makes a Puppy distro a Puppy, I would have to digress. I know from developing Carolina (and lina-lite, and also from observing the Saluki thread) that there are plenty of Puppy users out there that desire a modern back-end that can provide support for modern packages. There are plenty of users that have modern hardware that enjoy Puppy because it does provide a smaller footprint, not to mention a wonderful frugal install (which allows one to recover very quickly should they ‘trash’ their save-file); and speaking of a frugal install via a much more usable (and older) version of grub, it allows end-users to have multiple versions of Puppy side-by-side on their hard-drive in short order. No fuss, no muss. As far as Puppy support for older hardware goes, there already is a plethora of existing Puppy distros/derivatives that provide the Win95/WinXP era people with what they need. These are already available, rich in their diversity, thoroughly tested, and most of them very, very stable.

My main point in creating this thread was concerning the future of Puppy, in terms of keeping Puppy viable and attracting new users to the Puppy brand. Something like Puppy-Precise is at present very attractive, but what about down the road a bit? Ubuntu moves on. There’s no guarantee that woof2 will also move on. And I may be in the minority here, but I never cared for using another distro’s repositories; there’s too much lack of control in terms of pet size, and the said distro may make changes that could obsolete the Puppy backend, or particular packages, or even the Puppy version itself. Or, what if for example, down the road the Precise repository should cease to exist altogether? If no one is maintaining woof2, then where does that leave Puppy developers? To me, it would make more sense to put the woof2 emphasis on Debian support instead of Ubuntu. But even there it would need to be done in a manner that would incorporate the latest stable ‘flavor’ instead of hard-coding the present stable version. Things move on, and if woof2 depends on a hard-coded paradigm, then sometime down the road it will cease to be a viable method for creating Puppy distros. Actually, this is my main concern. That is the only reason that I suggested that Arch would be a better alternative because it provides a ‘rolling-release’. But all of those things could (and probably would) happen with Arch, too. And even though BarryK states that he’s successfully built Arch using woof2, and so have I, on my ‘puter it won’t go to a desktop. So there certainly are some issues there. I guess I’m trying to impress upon the Puppy forum that ‘woof2' isn’t even stable at present; what does that mean down the road in terms of Puppy development?

Yes, perhaps it would be simpler to “cut loose

simargl8

#10 Post by simargl8 »

Number of people in a team that makes Puppy derivative, or official release based Ubuntu is equal to zero. That puts Puppy to a list of dormant Linux distributions, or is that correct? No developers means no future releases, and that is not surprising, Puppy Linux was one man show, and only distribution without real package manager and build system.

Puppy shouldn't be based on T2 or LFS, it needs rewrite of pet package manager with support for build scripts like all other distributions, with proper dependencies handling, to be fast and reliable.

But, who will write new package manager and with what motivation, when he will never get even simple "Thank you"

User avatar
drongo
Posts: 374
Joined: Sat 10 Dec 2005, 23:35
Location: UK

ISO size

#11 Post by drongo »

Originally Puppy loaded everything into RAM. Later Puppies are more sophisticated.

So RAM size was the only criterion for limiting ISO size. If you can find Barry's original blog somewhere on the web this is all described in detail.

The fact that this made it easy to download over a low-speed link was an added bonus. (This was the reason I originally tried Puppy as I was downloading every small-sized ISO I could find in 2004).

nooby
Posts: 10369
Joined: Sun 29 Jun 2008, 19:05
Location: SwedenEurope

#12 Post by nooby »

I like what Raffy writes. I have EEPC and Acer D250 and D255
and an old Packard Bell with Celeron CPU and 512mb or was it 256MB RAM?

Ideal for lite distros like ?Puppy.

But am I really alone loving that Puppy do frugal install on NTFS
and same updates in save file in an easy to manually set up boot menu.lst

Only Porteus and latest Slack comes near in size and frugal install
on NTFS and with save file.
I use Google Search on Puppy Forum
not an ideal solution though

User avatar
mavrothal
Posts: 3096
Joined: Mon 24 Aug 2009, 18:23

#13 Post by mavrothal »

Linux distribution appear and disappear every year.
The ones that survive longer have a big core of developers (so can afford "defections" among other things).
Puppy is really a unique case of a 10-year old, one man show (and I do not mean to diminish all the people that contribute in puppy's development).
Puppy is what Barry Kauler wants it to be and so will be its future (he owns the name, you see).

The options I believe are
a) BK designating a (group of) person(s) to continue the Puppy/woof work
b) a (group of) person(s) working (together) along the woof lines will be christened by BK with "Puppy"
c) Puppy will follow the dormancy of other distros.
Given that BK (and Linux really) believes in do-ocracy, (b) is more likely that (a).

The real problem for (a) and (b) however, is the lack of either a financially secure, capable puppy developer that can devote 100% of his/her time on puppy (continuing the one-man-show model), or a group of developers with common ideas that can work together in a complementary way ( as this thread and many others suggest).

If these hold true for long, (c) will be the likely outcome. :(
But do not wary the real name of linux is "phoenix". It can always regenerate from its ashes (maybe as a different animal sometimes :D )
== [url=http://www.catb.org/esr/faqs/smart-questions.html]Here is how to solve your[/url] [url=https://www.chiark.greenend.org.uk/~sgtatham/bugs.html]Linux problems fast[/url] ==

nooby
Posts: 10369
Joined: Sun 29 Jun 2008, 19:05
Location: SwedenEurope

#14 Post by nooby »

And hardware tech moves on.

puppy are used to work on Intel x86 CPUs but them change too.

Intel latest power saving tech has not reached the Linux kernel yet?
Intel: Clover Trail Atom chips ‘cannot run Linux’

...
Clover Trail is an x86 chip and thus it will run any compliant code thrown at it.
Technically, there is no reason that a user can not wipe out Windows 8 and install a Linux distribution on the tablet.

The issue is that Clover Trail is packed to the brim with power saving features,
very few of which will work when using a Linux-based operating system.
...
so until Linux kernel devs catch up and include that code
it will run too hot or to have a poor battery life compared to Win 8.

but it is worse. Would Intel even give away that needed code for free
to the Linux devs? Not likely so only Ubuntu would have money enough
to pay for to get access to the power saving code needed.

http://www.extremetech.com/computing/13 ... -run-linux

that would suggest that maybe the niche for puppy linux is
to be a lite OS for older Intel Atom CPUs like the D250 and
the others like that and for the AMD cpu and a few more?

For to allow people to use Puppy on cheap available machines
everywhere using a CD or DVD or USB or doing frugal install
or wiping out the old windows if that is preferred.

To expect any dev to be so motivated and dedicated that they take on
proprietary Intel code and reverse engineer some patch seems not likely.
I use Google Search on Puppy Forum
not an ideal solution though

anikin
Posts: 994
Joined: Thu 10 May 2012, 06:16

#15 Post by anikin »

The way that article is written ... doesn't look to me as good reporting. Based on one particular processor model the author is making a far-reaching and broad-sweeping assumption. But he doesn't tell us, that Intel is the biggest contributor to Linux kernel. It has staff on its payroll, assigned to work exclusively on kernel development. Intel also supports the development of GCC. Starting from version 4.5.3, GCC has been steadily improving its support of Atom processors. Intel has a vested interest in Linux. It is a Linux friendly company, but the article implies otherwise.

darry1966

#16 Post by darry1966 »

The article with all due respect is from 2012 a lot has happened with the kernel since then.

nooby
Posts: 10369
Joined: Sun 29 Jun 2008, 19:05
Location: SwedenEurope

#17 Post by nooby »

It was the text that google found nearest in time
so other search terms is needed to find a more recent text.

Could you maybe give it a try to find such text?

Our wild guessing is not reliable. Seeing is believing.
I did a new search and September 2012 seems to be last time
anybody talked about "intel clover trail"

So what are we supposed to search for?

Linux intel power management
Now the firm is reported to have suffered delays in delivering power management software that will help Windows 8 take advantage of Clover Trail's power saving features.

With Intel heavily backing Windows 8 with its Clover Trail processor, this delay is a further embarrassment for the firm, coming just says after the company's CEO Paul Otellini told employees in Taiwan that Windows 8 will be shipped with bugs. According to Bloomberg, Microsoft has yet to approve any Clover Trail tablets due to Intel's delay in producing power management software.

Given that Intel has said that it won't be providing any power management software to the Linux kernel and seems unlikely to release the specifications of the Clover Trail chip, the chipmaker desperately needs Microsoft to green light Clover Trail tablet designs in time for the launch of Windows 8.

that was way back in October 2012 so we need something close in time.

a whole year ago. we need better key words for the search.
name of a chief linux kernel dev who takes care of intel power management maybe?
I use Google Search on Puppy Forum
not an ideal solution though

Dewbie

#18 Post by Dewbie »

raffy wrote:
Remember that MS even attempted to douse the enthusiasm for low-power "netbooks" (Atom-based mostly) by limiting its screen size and RAM?
A friend has an HP Mini, shipped with Win7 and 1G RAM; max RAM is 2G.
(Win7 will eventually need much more than that.)
I would love to see this strategy backfire, with users installing Linux en masse.

Post Reply