Mailing List Archive: Maemo: Developers

Why should it be so hard and should I even bother with Extras for fremantle?

Index | Next | Previous | View Threaded

thomas at mich

Oct 31, 2009, 7:10 PM

Post #1 of 24 (1061 views)
Permalink

Why should it be so hard and should I even bother with Extras for fremantle?

I'm a power user and not the only one. And what I used my current
tablets for were testing networks and doing other low level stuff,
mainly from xterm, but sometimes from python front-ends to linux. So
I ported a number of utilities under Linux to maemo/diablo and started
to do it for fremantle.

Way back when, I complained that half of what I wanted to install was
invisible except under "red pill mode", and after getting all the
noise about not using it and explanations, for those utilities I
thought were significant, I created versions in user/* because it is
the only way they would be visible.

One of them was socat. So I ported it and now that I submitted it for
fremantle they say it shouldn't be put in user/ so I'm both confused
and annoyed. This is the iTunes Application Store by mob justice. "I
don't like your app so you can't have it in extras!". No one reported
any actual bug, or problem, they just don't want to make it available
to users or anyone else using the normal means.

There is only one option and I'm trying to play by the rules - and
going thorough all the steps.

I have it in user/utilities which is probably where it belongs, but
someone suggested user/development. Normally I wouldn't mind, but
first, no guarantee anyone will actually change their vote, second, it
is really annoying since I have to bump the version too or it won't
build, reconstruct the source uploads, upload them, build them, hope
nothing goes wrong, and promote them JUST TO CHANGE ONE STUPID STRING
that someone doesn't like AND MAY NOT RESULT IN APPROVAL. A lot of
effort for probably nothing.

So we're back to having to do dpkg, hack around the application
manager, turn on or add "red pill mode" and all that junk again. Or
just tell everyone to enable extras-testing so they can have access to
disliked programs and have to put up with unstable betas?

Is there some category under user/* I can put it in without worrying
about being rejected except for actual bugs, or did all the discussion
about how to avoid the ugly hacks yet be able for users to actually
get to my programs result in nothing?
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

martin.grimme at gmail

Nov 1, 2009, 1:30 AM

Post #2 of 24 (1018 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

Hi,

it really looks wrong IMHO if stuff like socat, rootsh, or openssh
turn up in the extras repository ready for end-users to be installed.
extras should only contain applications that are safe for everyone to
play around with.
Stuff like this is meant for advanced users and developers. I think it
would be best to have another official repository "extras-advanced"
for these things, that comes preconfigured but deactivated on the
device, with a big warning that apps in there are meant for advanced
users who know what these tools are.

Maybe we should open a brainstorm for separating the sharp knives from
the toys. :)

extras-advanced could have its own QA queue even. Or the QA decides
whether something that passes QA is good for extras (safe for
everyone) or extras-advanced (tools for advanced users). Maybe
something like voting with "thumbs up (end-user)", "thumbs up
(advanced)", and "thumbs down"?

Any suggestions?

Martin

2009/11/1, tz <thomas [at] mich>:
> I'm a power user and not the only one. And what I used my current
> tablets for were testing networks and doing other low level stuff,
> mainly from xterm, but sometimes from python front-ends to linux. So
> I ported a number of utilities under Linux to maemo/diablo and started
> to do it for fremantle.
>
> Way back when, I complained that half of what I wanted to install was
> invisible except under "red pill mode", and after getting all the
> noise about not using it and explanations, for those utilities I
> thought were significant, I created versions in user/* because it is
> the only way they would be visible.
>
> One of them was socat. So I ported it and now that I submitted it for
> fremantle they say it shouldn't be put in user/ so I'm both confused
> and annoyed. This is the iTunes Application Store by mob justice. "I
> don't like your app so you can't have it in extras!". No one reported
> any actual bug, or problem, they just don't want to make it available
> to users or anyone else using the normal means.
>
> There is only one option and I'm trying to play by the rules - and
> going thorough all the steps.
>
> I have it in user/utilities which is probably where it belongs, but
> someone suggested user/development. Normally I wouldn't mind, but
> first, no guarantee anyone will actually change their vote, second, it
> is really annoying since I have to bump the version too or it won't
> build, reconstruct the source uploads, upload them, build them, hope
> nothing goes wrong, and promote them JUST TO CHANGE ONE STUPID STRING
> that someone doesn't like AND MAY NOT RESULT IN APPROVAL. A lot of
> effort for probably nothing.
>
> So we're back to having to do dpkg, hack around the application
> manager, turn on or add "red pill mode" and all that junk again. Or
> just tell everyone to enable extras-testing so they can have access to
> disliked programs and have to put up with unstable betas?
>
> Is there some category under user/* I can put it in without worrying
> about being rejected except for actual bugs, or did all the discussion
> about how to avoid the ugly hacks yet be able for users to actually
> get to my programs result in nothing?
> _______________________________________________
> maemo-developers mailing list
> maemo-developers [at] maemo
> https://lists.maemo.org/mailman/listinfo/maemo-developers
>
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

david at dgreaves

Nov 1, 2009, 1:46 AM

Post #3 of 24 (1017 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sat, 2009-10-31 at 19:10 -0700, tz wrote:
> I'm a power user and not the only one.
Agreed.

> And what I used my current
> tablets for were testing networks and doing other low level stuff,
> mainly from xterm, but sometimes from python front-ends to linux. So
> I ported a number of utilities under Linux to maemo/diablo and started
> to do it for fremantle.

And this is the problem.

If you got all the people in the world who want to use the tablet/phone
in this manner you wouldn't have a market that warranted making the
device. Personally I want the thing to be uber-successful and I want to
be on the inside where I can ensure *my* niche needs are met by working
collaboratively with those who have to support the masses.

Otherwise good luck writing socat for a nailed down iPhone (in a year or
few).

> Way back when, I complained that half of what I wanted to install was
> invisible except under "red pill mode", and after getting all the
> noise about not using it and explanations, for those utilities I
> thought were significant, I created versions in user/* because it is
> the only way they would be visible.
> One of them was socat. So I ported it and now that I submitted it for
> fremantle they say it shouldn't be put in user/ so I'm both confused
> and annoyed.
Really?
Because you think that 'socat' is something that a wide cross-section of
the target audience will use?

> This is the iTunes Application Store by mob justice.
> "I don't like your app so you can't have it in extras!".
ROFL.

> No one reported
> any actual bug, or problem, they just don't want to make it available
> to users or anyone else using the normal means.
However, that's a valid point.

The bug "Way too geeky to present to most users" should, IMHO be an
'extras user/*' criteria.

> There is only one option and I'm trying to play by the rules - and
> going thorough all the steps.

<snip>
rant about turmoil when involved in the early phases of something and
identifying a bug in a process.
</snip>

> Is there some category under user/* I can put it in without worrying
> about being rejected except for actual bugs, or did all the discussion
> about how to avoid the ugly hacks yet be able for users to actually
> get to my programs result in nothing?

Ah, the actual point :)
(and a good one too).

Alongside 'user/*' I wonder if we should have a 'geek/*' section ?
Or make 'user/development' and some other categories only visible if
enabled in preferences.

Personally I think we're back at the Categories argument that was never
really taken seriously.

David

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

andrew at bleb

Nov 1, 2009, 2:22 AM

Post #4 of 24 (1021 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sun, Nov 1, 2009 at 09:30, Martin Grimme <martin.grimme [at] gmail> wrote:
>
> it really looks wrong IMHO if stuff like socat, rootsh, or openssh
> turn up in the extras repository ready for end-users to be installed.
> extras should only contain applications that are safe for everyone to
> play around with.

However, if it is clear to users what to expect when they install
those apps; and they don't recklessly reduce the capability of the
system, I don't think they're that bad. So, perhaps, OpenSSH Server
should have a clearer warning that the root password it forces you to
choose should be strong, as it will allow anyone to log in to your
tablet if it is online and they know the password.

> Stuff like this is meant for advanced users and developers. I think it
> would be best to have another official repository "extras-advanced"
> for these things, that comes preconfigured but deactivated on the
> device, with a big warning that apps in there are meant for advanced
> users who know what these tools are.

It seems like a lot of work. In the socat case, as long as it has a
nice icon and a description & summary which makes it clear that this
is a command-line tool for usage in X Terminal - and adequately
explains socat itself; I think it is a fine fit for user/utilities or,
even better, user/network[1].

However, the process is still being worked out. And, as I'm about to
outline, I think Thomas was right in bringing this to a mailing list
rather than having a debate either one-on-one or on the package page
(althought perhaps in a slightly more constructive way, no one wants
to prevent useful stuff getting to people; the process is still being
worked out and hopefully Thomas wouldn't think that users should have
a positive experience). The mailing list can act as arbiter.

Hope that helps,

Andrew

[1] http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging%2C_Deploying_and_Distributing#Sections

<off-topic>
> Maybe we should open a brainstorm for separating the sharp knives from
> the toys. :)

Can anyone point to a brainstorm idea which has been successfully
implemented, or produced an outcome which reflects the benefits of an
actual brainstorm? I'm increasingly forming the opinion that
brainstorm users often conflate problems and solutions, and the whole
concept is based around the opinion that everyone's opinion and
solutions are equal. They aren't. Some people design better systems
than others.

Similarly, there's no chance of persuasion and debate. Mailing lists
are a much more powerful tool, IMHO. Perhaps brainstorm is best to eke
out the *requirements*, which are then fed to the experts on a mailing
list for debate and discussion as to how best to implement those
requirements? (Not that brainstorm really results in any ownership
either)

However, if people can point to it as a success story, I'll be very
happy to be dissuaded of this notion :-)
</off-topic>

--
Andrew Flegg -- mailto:andrew [at] bleb | http://www.bleb.org/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

g+770 at cobb

Nov 1, 2009, 3:35 AM

Post #5 of 24 (1028 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sunday 01 November 2009 10:22:00 Andrew Flegg wrote:
> On Sun, Nov 1, 2009 at 09:30, Martin Grimme <martin.grimme [at] gmail> wrote:
> > it really looks wrong IMHO if stuff like socat, rootsh, or openssh
> > turn up in the extras repository ready for end-users to be installed.
> > extras should only contain applications that are safe for everyone to
> > play around with.
>
> However, if it is clear to users what to expect when they install
> those apps; and they don't recklessly reduce the capability of the
> system, I don't think they're that bad. So, perhaps, OpenSSH Server
> should have a clearer warning that the root password it forces you to
> choose should be strong, as it will allow anyone to log in to your
> tablet if it is online and they know the password.

I agree with Andrew. These applications should be available to everyone.
There are two issues:

1) There will end up being a LOT of command line utilities. Over time I would
expect a lot of debian utilities to be ported. This will clutter up the
Application Manager with things of interest to a tiny number of people.

The medium term solution, I think, is to create a new category: user/advanced
or user/command-line.

In the short term, I suggest we look at creating a user/utilities meta-package
called something like "Command Line Tools" which would depend on the (non
user/) tools themselves. Whenever someone ports one of these tools they
would work with the maintainer of the "Command Line Tools" package to have it
added to the Description and to the dependencies. We might even manually
force the promotion of the new "Command Line Tools" package if someone
(X-Fade?) is happy that it is just a trivial port.

2) Some of the tools may be fairly dangerous. But then that is true of many
of the tools already on the system (mke2fs?). I don't see that as a problem
if it is marked as an advanced tool except for cases (like OpenSSH Server)
where just installing it does something potentially dangerous to your system
(another example might be something which installs a daemon which does
something very useful but which reduces battery life).

In this case, the medium term solution is, again, to put it in a user/advanced
category and the short term solution is, I think, just to make it explain how
it is dangerous in BIG LETTERS!

Graham

P.S. Can we please start a new thread for discussion of the brainstorm
process?
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

andrew at bleb

Nov 1, 2009, 3:55 AM

Post #6 of 24 (1031 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sun, Nov 1, 2009 at 11:35, Graham Cobb <g+770 [at] cobb> wrote:
> On Sunday 01 November 2009 10:22:00 Andrew Flegg wrote:
>>
>> However, if it is clear to users what to expect when they install
>> those apps; and they don't recklessly reduce the capability of the
>> system, I don't think they're that bad. So, perhaps, OpenSSH Server
>> should have a clearer warning that the root password it forces you to
>> choose should be strong, as it will allow anyone to log in to your
>> tablet if it is online and they know the password.
>
> I agree with Andrew. �These applications should be available to everyone.

Cool.

> 1) There will end up being a LOT of command line utilities. �Over time I would
> expect a lot of debian utilities to be ported. �This will clutter up the
> Application Manager with things of interest to a tiny number of people.
>
> The medium term solution, I think, is to create a new category: user/advanced
> or user/command-line.

I don't like this. If I'm looking for a port forwarder, I'm going to
look in user/network and, as an advanced user, I'm going to be happy
with a pretty GUI app or the command-line socat.

This is probably where we should start looking at introducing debtags?
So, for example, socat would be in user/network but have a tag,
'command-line'. We can then even add a setting to Application Manager
to show/hide command-line applications.

This set-up then is an obvious first step on the wider use of debtags,
which will garner us essential experience in how best to use them. (I
believe it's important that if the user selects one category they see
one set of apps. I believe it's important that one app appears in one
category: otherwise the user might not be clear that they're
successfully navigating from one category to the next if they see the
same apps cropping up over and over again. I also believe that tags
could make authors lazy in correctly categorising their applications)

I've also got a vested interest as a maintainer of the vim port, but
then there's also an application menu item for it which launches it
within osso-xterm. Yes, it's "command line" but it's also got a (very
basic) GUI. If there *was* a command-line tag switch, would I include
the tag?

> In the short term, I suggest we look at creating a user/utilities meta-package
> called something like "Command Line Tools" which would depend on the (non
> user/) tools themselves.

Hmm, seems messy. But perhaps that's a simple solution. I'm not sure
it'll work in practice unless the maintainer(s) of the "Command Line
Tools" package are quick to respond to change.

Depends on how many purely command-line utils (rather than full-blown
apps which can have an osso-xterm launched from the menu) we'll
*actually* see.

> P.S. Can we please start a new thread for discussion of the brainstorm
> process?

Done. Taken to -community.

Cheers,

Andrew

--
Andrew Flegg -- mailto:andrew [at] bleb | http://www.bleb.org/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

anidel at gmail

Nov 1, 2009, 6:00 AM

Post #7 of 24 (1023 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

2009/11/1 Andrew Flegg <andrew [at] bleb>:
> On Sun, Nov 1, 2009 at 11:35, Graham Cobb <g+770 [at] cobb> wrote:
>> On Sunday 01 November 2009 10:22:00 Andrew Flegg wrote:
>>>
>>> However, if it is clear to users what to expect when they install
>>> those apps; and they don't recklessly reduce the capability of the
>>> system, I don't think they're that bad. So, perhaps, OpenSSH Server
>>> should have a clearer warning that the root password it forces you to
>>> choose should be strong, as it will allow anyone to log in to your
>>> tablet if it is online and they know the password.
>>
>> I agree with Andrew. �These applications should be available to everyone.
>
> Cool.
>

+1 (for the agreement part, not the coolness :) )

>> 1) There will end up being a LOT of command line utilities. �Over time I would
>> expect a lot of debian utilities to be ported. �This will clutter up the
>> Application Manager with things of interest to a tiny number of people.
>>
>> The medium term solution, I think, is to create a new category: user/advanced
>> or user/command-line.
>
> I don't like this. If I'm looking for a port forwarder, I'm going to
> look in user/network and, as an advanced user, I'm going to be happy
> with a pretty GUI app or the command-line socat.
>
> This is probably where we should start looking at introducing debtags?
> So, for example, socat would be in user/network but have a tag,
> 'command-line'. We can then even add a setting to Application Manager
> to show/hide command-line applications.
>

I would like to see this.
'socat' definitely belongs to user/network but is also a CLI utility (advanced).

--
anidel
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

pav at iki

Nov 1, 2009, 11:21 AM

Post #8 of 24 (1008 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

su, 2009-11-01 kello 09:46 +0000, David Greaves kirjoitti:
[clip]
> Alongside 'user/*' I wonder if we should have a 'geek/*' section ?
> Or make 'user/development' and some other categories only visible if
> enabled in preferences.
>
> Personally I think we're back at the Categories argument that was never
> really taken seriously.

A separate tree of categories could be useful. Though maybe it would be
better to name it 'expert/*', and warn in the UI that these are expert
tools that really may require expertise to use properly or safely.

What would it take to have the application manager to have a setting
"Show expert tools" that would show/hide these? Or maybe have these only
installable from the command line?

--
Pauli Virtanen

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

andrew at bleb

Nov 1, 2009, 11:39 AM

Post #9 of 24 (1016 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sun, Nov 1, 2009 at 19:21, Pauli Virtanen <pav [at] iki> wrote:
>
> What would it take to have the application manager to have a setting
> "Show expert tools" that would show/hide these?

I'm slightly familiar with the HAM code, it should be fairly trivial.
Getting it into users' hands is slightly trickier, mind you.

> Or maybe have these only installable from the command line?

Then there's no benefit to having them visible in HAM. The tool which
started this discussion, socat, could easily be installed by someone
doing `apt-get install socat' as root.

This logic was part of the original rationale for having no 'cli'
section (apart from the fact it's *totally* unorthogonal): users who
could run them could use apt-get and apt-cache just as easily as the
tools themselves.

Cheers,

Andrew

--
Andrew Flegg -- mailto:andrew [at] bleb | http://www.bleb.org/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

jeremiah at jeremiahfoster

Nov 1, 2009, 3:32 PM

Post #10 of 24 (1002 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Nov 1, 2009, at 12:35, Graham Cobb wrote:

> On Sunday 01 November 2009 10:22:00 Andrew Flegg wrote:
>> On Sun, Nov 1, 2009 at 09:30, Martin Grimme
>> <martin.grimme [at] gmail> wrote:
>>> it really looks wrong IMHO if stuff like socat, rootsh, or openssh
>>> turn up in the extras repository ready for end-users to be
>>> installed.
>>> extras should only contain applications that are safe for everyone
>>> to
>>> play around with.
>>
>> However, if it is clear to users what to expect when they install
>> those apps; and they don't recklessly reduce the capability of the
>> system, I don't think they're that bad. So, perhaps, OpenSSH Server
>> should have a clearer warning that the root password it forces you to
>> choose should be strong, as it will allow anyone to log in to your
>> tablet if it is online and they know the password.
>
> I agree with Andrew. These applications should be available to
> everyone.
> There are two issues:
>
> 1) There will end up being a LOT of command line utilities. Over
> time I would
> expect a lot of debian utilities to be ported. This will clutter up
> the
> Application Manager with things of interest to a tiny number of
> people.

Then Application Manager has to change. It does not scale to have
categories when there are thousands of apps.
>
> The medium term solution, I think, is to create a new category: user/
> advanced
> or user/command-line.

I'm not so keen on this. I think it excessively cautious.
>
> In the short term, I suggest we look at creating a user/utilities
> meta-package
> called something like "Command Line Tools" which would depend on the
> (non
> user/) tools themselves. Whenever someone ports one of these tools
> they
> would work with the maintainer of the "Command Line Tools" package
> to have it
> added to the Description and to the dependencies. We might even
> manually
> force the promotion of the new "Command Line Tools" package if someone
> (X-Fade?) is happy that it is just a trivial port.
>
> 2) Some of the tools may be fairly dangerous. But then that is true
> of many
> of the tools already on the system (mke2fs?). I don't see that as a
> problem
> if it is marked as an advanced tool except for cases (like OpenSSH
> Server)
> where just installing it does something potentially dangerous to
> your system
> (another example might be something which installs a daemon which does
> something very useful but which reduces battery life).
>
> In this case, the medium term solution is, again, to put it in a
> user/advanced
> category and the short term solution is, I think, just to make it
> explain how
> it is dangerous in BIG LETTERS!

UNIX, and Linux, has always been a system in which you can do
significant damage to your system by issuing the wrong command. I
don't think you can dumb it down and still keep its power and
flexibility. I realize that there are interfaces to GSM and other data
that users oughtn't touch, but I think the warnings and safeguards we
have are sufficient.

Jeremiah

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

khertan at khertan

Nov 1, 2009, 9:47 PM

Post #11 of 24 (991 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

Are you all kidding ? seriously ?

"The bug "Way too geeky to present to most users" should, IMHO be an
'extras user/*' criteria."

Does we need to come back to the old days where each user have his own
repository ? To be honest i ll be far away easier to do than the
actual one.

2009/11/2 Jeremiah Foster <jeremiah [at] jeremiahfoster>:
>
>
> Begin forwarded message:
>
>> From: Jeremiah Foster <jeremiah [at] jeremiahfoster>
>> Date: November 2, 2009 12:35:45 AM GMT+01:00
>> To: Andrew Flegg <andrew [at] bleb>
>> Subject: Re: Why should it be so hard and should I even bother with
>> Extras for fremantle?
>>
>>
>> On Nov 1, 2009, at 12:55, Andrew Flegg wrote:
>>
>>> On Sun, Nov 1, 2009 at 11:35, Graham Cobb <g+770 [at] cobb> wrote:
>>>> On Sunday 01 November 2009 10:22:00 Andrew Flegg wrote:
>>>
>>>> 1) There will end up being a LOT of command line utilities. Over
>>>> time I would
>>>> expect a lot of debian utilities to be ported. This will clutter
>>>> up the
>>>> Application Manager with things of interest to a tiny number of
>>>> people.
>>>>
>>>> The medium term solution, I think, is to create a new category:
>>>> user/advanced
>>>> or user/command-line.
>>>
>>> I don't like this. If I'm looking for a port forwarder, I'm going to
>>> look in user/network and, as an advanced user, I'm going to be happy
>>> with a pretty GUI app or the command-line socat.
>>>
>>> This is probably where we should start looking at introducing
>>> debtags?
>>> So, for example, socat would be in user/network but have a tag,
>>> 'command-line'. We can then even add a setting to Application Manager
>>> to show/hide command-line applications.
>>
>> This makes lots of sense and has been shown to work, although not
>> every debian app has tags.
>>>
>>> This set-up then is an obvious first step on the wider use of
>>> debtags,
>>> which will garner us essential experience in how best to use them. (I
>>> believe it's important that if the user selects one category they see
>>> one set of apps. I believe it's important that one app appears in one
>>> category: otherwise the user might not be clear that they're
>>> successfully navigating from one category to the next if they see the
>>> same apps cropping up over and over again. I also believe that tags
>>> could make authors lazy in correctly categorising their applications)
>>
>> Power users are often a huge help here because they can categorize
>> themselves as well.
>>>
>>> I've also got a vested interest as a maintainer of the vim port,
>>
>> Which I would tag as: not::emacs modal::editor donot::use
>>
>> Jeremiah
>
> _______________________________________________
> maemo-developers mailing list
> maemo-developers [at] maemo
> https://lists.maemo.org/mailman/listinfo/maemo-developers
>

--
Benoît HERVIER - http://khertan.net/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

matan at svgalib

Nov 1, 2009, 11:20 PM

Post #12 of 24 (993 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Sun, 1 Nov 2009, Andrew Flegg wrote:

> Then there's no benefit to having them visible in HAM. The tool which
> started this discussion, socat, could easily be installed by someone
> doing `apt-get install socat' as root.

Do you know that almost any current Linux distribution has a GUI package
installation program? And that each such program displays thousands of
CLI and advanced GUI programs? Are you sure that reinventing this wheel
is a good idea?

--
Matan.

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

andrew at bleb

Nov 1, 2009, 11:52 PM

Post #13 of 24 (998 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

Benoît, you wrote:
>
> Are you all kidding ? seriously ?
>
> "The bug "Way too geeky to present to most users" should, IMHO be an
> 'extras user/*' criteria."
>
> Does we need to come back to the old days where each user have his own
> repository ? To be honest i ll be far away easier to do than the
> actual one.

You seem to be replying to Jeremiah, replying to me, replying to Graham and none of the three of us you've quoted are suggesting that such apps are excluded from Extras.

Were you just piggy-backing on the most recent email?

Cheers,

Andrew

--
Andrew Flegg -- mailto:andrew [at] bleb | http://www.bleb.org/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

marius.vollmer at nokia

Nov 2, 2009, 2:47 AM

Post #14 of 24 (986 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

ext Jeremiah Foster <jeremiah [at] jeremiahfoster> writes:

> Then Application Manager has to change. It does not scale to have
> categories when there are thousands of apps.

Yes, this is overdue. We also need a way to handle application specific
add-ons, like language packs.
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

andrew at bleb

Nov 2, 2009, 2:51 AM

Post #15 of 24 (995 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Mon, Nov 2, 2009 at 10:47, Marius Vollmer <marius.vollmer [at] nokia> wrote:
> ext Jeremiah Foster <jeremiah [at] jeremiahfoster> writes:
>
>> Then Application Manager has to change. It does not scale to have
>> categories when there are thousands of apps.
>
> Yes, this is overdue. �We also need a way to handle application specific
> add-ons, like language packs.

Instead of debtags, the subsection approach of
user/network/command-line (or user/network/advanced) could solve both
problems in one.

Cheers,

Andrew

--
Andrew Flegg -- mailto:andrew [at] bleb | http://www.bleb.org/
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

marius.vollmer at nokia

Nov 2, 2009, 2:53 AM

Post #16 of 24 (987 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

ext Pauli Virtanen <pav [at] iki> writes:

> A separate tree of categories could be useful. Though maybe it would be
> better to name it 'expert/*', and warn in the UI that these are expert
> tools that really may require expertise to use properly or safely.
>
> What would it take to have the application manager to have a setting
> "Show expert tools" that would show/hide these? Or maybe have these only
> installable from the command line?

This would work for beta versions of software, too. There could be a
"Show beta versions" button. Or better: hierachical categories.

As someone mentioned, we should probably start looking at debtags to
encode all this instead of using the "Section:" field.
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

pav+sp at iki

Nov 2, 2009, 3:22 AM

Post #17 of 24 (982 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

Mon, 02 Nov 2009 09:20:41 +0200, Matan Ziv-Av wrote:
> On Sun, 1 Nov 2009, Andrew Flegg wrote:
>
>> Then there's no benefit to having them visible in HAM. The tool which
>> started this discussion, socat, could easily be installed by someone
>> doing `apt-get install socat' as root.
>
> Do you know that almost any current Linux distribution has a GUI package
> installation program? And that each such program displays thousands of
> CLI and advanced GUI programs? Are you sure that reinventing this wheel
> is a good idea?

This is largely the current situation. But I think it's unwise to say
that it's the optimal solution, so of course trying to improve this is a
good idea.

In fact, one could say that the trend in Linux distributions is to move
away from this. Consider Ubuntu: they have a separate "Software center"
that shows only a subset of the available software and aims to be simple
to use.

--
Pauli Virtanen

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

dieter at plaetinck

Nov 2, 2009, 3:29 AM

Post #18 of 24 (1001 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Mon, 2 Nov 2009 10:51:25 +0000
Andrew Flegg <andrew [at] bleb> wrote:

> On Mon, Nov 2, 2009 at 10:47, Marius Vollmer
> <marius.vollmer [at] nokia> wrote:
> > ext Jeremiah Foster <jeremiah [at] jeremiahfoster> writes:
> >
> >> Then Application Manager has to change. It does not scale to have
> >> categories when there are thousands of apps.
> >
> > Yes, this is overdue. We also need a way to handle application
> > specific add-ons, like language packs.
>
> Instead of debtags, the subsection approach of
> user/network/command-line (or user/network/advanced) could solve both
> problems in one.
>
> Cheers,
>
> Andrew
>

usually tags are a better approach then hierachical trees because often
items belong in multiple categories.

and UI-wise, you can do very nice things with tags (toggle one or more
tags you want to see, or "all/current selection minus selected tags",
though it maybe a bit more complicated to find apps that would
otherwise be in the rootlevel.

Dieter
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

david at dgreaves

Nov 2, 2009, 12:17 PM

Post #19 of 24 (980 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Mon, 2009-11-02 at 06:47 +0100, Benoît HERVIER wrote:
> Are you all kidding ? seriously ?
> "The bug "Way too geeky to present to most users" should, IMHO be an
> 'extras user/*' criteria."
I am serious.

For goodness sake.... didn't you spot that the device ships with a
Facebook applet? If the device is to succeed then even people who
believe horoscopes must be comfortable using it!

"Accidentally" presenting them with a network sniffer application is,
IMHO, a user experience bug.

> Does we need to come back to the old days where each user have his own
> repository ? To be honest i ll be far away easier to do than the
> actual one.

Well, yes - and clearly NO.

But through the marvel of views and filters we can tailor a single
repository to multiple kinds of user.

Maybe have the HAM run in 3 views:
casual : games, social, media, music
professional: spreadsheets, sync apps
technical: network sniffers, encryption tools, CLI

David

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

jeremiah at jeremiahfoster

Nov 2, 2009, 2:19 PM

Post #20 of 24 (980 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Nov 2, 2009, at 21:17, David Greaves wrote:

> If the device is to succeed then even people who
> believe horoscopes must be comfortable using it!

You must be a taurus . . .
>
> "Accidentally" presenting them with a network sniffer application is,
> IMHO, a user experience bug.
>
>> Does we need to come back to the old days where each user have his
>> own
>> repository ? To be honest i ll be far away easier to do than the
>> actual one.
>
> Well, yes - and clearly NO.
>
> But through the marvel of views and filters we can tailor a single
> repository to multiple kinds of user.
>
> Maybe have the HAM run in 3 views:
> casual : games, social, media, music
> professional: spreadsheets, sync apps
> technical: network sniffers, encryption tools, CLI

I think the two suggestions on this thread that are present in
upstream and have been shown to work are:

- The implementation of tags
- The simplification of user interface for installing apps

It does have to be really, really easy to get a twitter client
installed via a few clicks in a GUI. And it also has to be really,
really easy for a Qt python hacker to find the best lib to create that
twitter client. I think these are separate domains which require
separate tools. Perhaps we can simplify the Application Manager and
trust that developers will more easily find libs and such through the
implementation of tags in the repos?

Jeremiah
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

thomas at mich

Nov 4, 2009, 4:09 PM

Post #21 of 24 (889 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

To extend or clarify another issue:

There are things which should be and are in other sections like python
and associated libs. Lots of things like this and cli stuff are
building-blocks, i.e. I might want to use a maintained and tested
utility for several things and which others can use instead of copying
a routine and having it duplicated.

Karma would work for this but not for user/* - if I update something
outside of user and in something which won't be visible in application
manager, how can I get it to users without a user stub whose only
purpose is to pull in the updated library?

On Wed, Nov 4, 2009 at 4:04 PM, tz <thomas [at] mich> wrote:
> I was hoping to spark a discussion, and I think it worked out better
> than I was expecting for a rant.
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

kees.jongenburger at gmail

Nov 5, 2009, 1:34 AM

Post #22 of 24 (876 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

Hi,

On Mon, Nov 2, 2009 at 11:51 AM, Andrew Flegg <andrew [at] bleb> wrote:
> On Mon, Nov 2, 2009 at 10:47, Marius Vollmer <marius.vollmer [at] nokia> wrote:
>> ext Jeremiah Foster <jeremiah [at] jeremiahfoster> writes:
>>
>>> Then Application Manager has to change. It does not scale to have
>>> categories when there are thousands of apps.
>>
>> Yes, this is overdue. �We also need a way to handle application specific
>> add-ons, like language packs.
>
> Instead of debtags, the subsection approach of
> user/network/command-line (or user/network/advanced) could solve both
> problems in one.

How about a totally different approach. Pretty much like the currently
nokia-binaries have a unique url
would it be possible to create "per" user repositories on the
maemo.org server and have some tool virtually merging
different repositories into this personal repository.

The rationale would be that you could do most stuff "online" and that
you can make your "application installer" even more simple. not
worrying about
all those sections etc. You could also possibly reflash your device
and get all the apps you had back.

Greetings
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

maemo at csipa

Nov 5, 2009, 1:54 AM

Post #23 of 24 (890 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Thursday 05 November 2009 10:34:53 Kees Jongenburger wrote:
> How about a totally different approach. Pretty much like the currently
> nokia-binaries have a unique url
> would it be possible to create "per" user repositories on the
> maemo.org server and have some tool virtually merging
> different repositories into this personal repository.

Techinally, you can do this through /etc/apt/preferences if I understand
correctly what you want to do :)

BTW not to sidetrack, but per user repositories could be used as PPAs. I
understand the motives of having a monolithic extras-devel, but in some cases
it's counterproductive. If I want to make separate packages (for comparison)
or want to try out some super-dangerous stuff with a limited number of people,
I don't want to do it in extras-devel (you might pull in my package as a
dependency by accident, cleaning up extras-devel from test versions also
involves poking people...).

_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

jeremiah at jeremiahfoster

Nov 5, 2009, 9:04 AM

Post #24 of 24 (865 views)
Permalink

Re: Why should it be so hard and should I even bother with Extras for fremantle? [In reply to]

On Nov 5, 2009, at 10:54, Attila Csipa wrote:

> On Thursday 05 November 2009 10:34:53 Kees Jongenburger wrote:
>> How about a totally different approach. Pretty much like the
>> currently
>> nokia-binaries have a unique url
>> would it be possible to create "per" user repositories on the
>> maemo.org server and have some tool virtually merging
>> different repositories into this personal repository.
>
> Techinally, you can do this through /etc/apt/preferences if I
> understand
> correctly what you want to do :)
>
> BTW not to sidetrack, but per user repositories could be used as
> PPAs. I
> understand the motives of having a monolithic extras-devel, but in
> some cases
> it's counterproductive. If I want to make separate packages (for
> comparison)
> or want to try out some super-dangerous stuff with a limited number
> of people,
> I don't want to do it in extras-devel (you might pull in my package
> as a
> dependency by accident, cleaning up extras-devel from test versions
> also
> involves poking people...).

I'll add PPAs or personal user repos as an item for disucssion.

Jeremiah
_______________________________________________
maemo-developers mailing list
maemo-developers [at] maemo
https://lists.maemo.org/mailman/listinfo/maemo-developers

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads