replacing the Startup Disk Creator

[Nio Wiklund]

Den 2015-09-17 kl. 19:08, skrev Mathieu Trudel-Lapierre:
> Le 2015-09-17 12:44, Nio Wiklund a écrit :
> [...]
>> I will forward your comments and tips to the persons at the Ubuntu
>> Forums who reported the bugs and bad features :-)
>>
>> More than four years has passed with severe bugs, that make the SDC fail
>> in many cases. I think you have had enough chances to fix it, and it
>> would be better to use another tool instead. That way you can spend your
>> time improving other things.
>>
>> I can only suggest a change, and supply information and opinion about
>> the alternative tools. If a majority of the deciding forum does not want
>> to replace the SDC, that's it. We can continue to help people at the
>> Ubuntu Forums to install other tools in order to make USB pendrives
>> bootable with Ubuntu.
>>
> 
> I've seen the forum survey on this matter already. The trouble is,
> unetbootin is complicated (does it do persistence correctly?).
> multisystem (although I am not familiar with it) is likely to be
> complicated enough too given it's purpose. I don't know about mkusb, but
> the UI looks clunky at first glance.
> 
> Disks requires too many clicks to reach the right interface, and
> therefore would need documentation for new users, although it seems like
> it could well do the trick (also, doesn't do persistence).
> 
> And let's not talk about dd, which is very dangerous for a new user to
> use. It's far too likely for someone to nuke their hard drive using it.
> 
> These are just the top options from the poll, with 31 voters. It's not
> nearly representative enough to make a decision.
> 
> In the end, we're already using usb-creator, and maintaining it, and the
> code itself is very approachable. I have no idea who would volunteer to
> maintain an alternative to usb-creator either. I think the price of
> replacing it for the moment far outweighs the cost of fixing the few
> bugs it has to make it work well for everyone.
> 
> As for a timeline, I would like to propose something:
> 
> It's far too late this cycle to change tools, and it seems ambitious to
> me to change it for a LTS when it mostly works -- let's try to fix the
> bugs between now and the next LTS, and revisit afterwards if there is a
> need to change. In the meantime, this will allow time for a proper
> proposal to be made, and for any fixes to be done for an alternative to
> be rock solid. I'm reasonably certain that if everyone pitches in to
> report the bugs and propose patches if they can, we can fix the top
> issues easily.

Hi again Mathieu,

It seems that a fair fraction of those who have bothered to vote in the
poll and / or written a detailed reply prefer Unetbootin. And I know
that many of them have used it in many situations.

There have been problems with it, for example the gfxboot.c32: not a
COM32R Image boot bug (corresponding to #1325801 for the SDC). But it
was quickly fixed by the developer and a working version was available
at the developer's PPA. Unfortunately the version of Unetbootin in the
Ubuntu repository is lagging behind.

Persistence works for me, when created with Unetbootin. It is provided
with a file, like it is by the SDC. I provide a casper-rw partition
(ext4 without journaling) with mkusb. I also provide a persistent live
system that works in both UEFI and BIOS mode (even with 32-bit Ubuntu
systems). But the main task of mkusb is 'wrap a safety belt around dd',
and the dd process is extremely simple and reliable for a great variety
of linux distros (when you point to the correct target device).

If it is only a question of improving the interface, I'm willing to do
it in cooperation with whoever is available from your developing group.
I could also re-write it in a simpler way (and peel off some special
features, for example the shortcut for isotesting).

Disks (gnome-disks) is also very simple and reliable. A slightly
modified interface, and advertizing it for this purpose (to make USB
boot drives) would make it a very good alternative to replace the SDC.
That would take care of at least 90% of the usage.

The remainder, when people want persistence or use the rest of the drive
space for carrying data between computers, or who want a multiboot
pendrive, can be managed with Unetbootin or Multisystem, but also with
some other tools that 'participate' in the poll at the Ubuntu Forums.
LXLE replaced the SDC with MultibootUSB in the current release. They
have tested several tools before selecting it. ToriOS will soon release
its first official version, and it uses mkusb.

-o-

I understand that your intentions with the SDC are good, but the record
from the four last years tells another story. It does not work well,
most people who know other tools prefer one of them (at least this is my
experience from the Ubuntu Forums). There are different favourites, but
at the end of our poll (late October), I think that Unetbootin will have
the highest score by a wide margin.

'I think the price of replacing it for the moment far outweighs the cost
of fixing the few bugs it has to make it work well for everyone.'

You must also count the 'cost' of losing users, because the built-in
tool does not work well.

-o-

The alternative tools have developers/maintainers already. I think you
will need someone who makes sure that it continues to work with new
versions and across version boundaries in Ubuntu. Depending on what you
choose, that task can be easier or more difficult. But as it is now,
those alternatives already work better than the SDC, so I would expect
the task to be fairly easy, if the persons can cooperate.

I understand that it is too late for Wily, but it should be possible to
replace the SDC with a tool that already works better in the next
version. It happens to be an LTS, and it would be great to be able to
offer such an improvement in our next LTS release :-)

Best regards
Nio