[PATCH] gtk-gui.c: Move the download dialog related code to a new file

[Dirk Hohndel]

Linus Torvalds <torvalds at linux-foundation.org> writes:

> On Thu, Jan 10, 2013 at 12:06 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
>>
>> The problem is that while you are a user of Subsurface I don't think you
>> are the typical user. Most people will use Subsurface on one system with
>> one (or a few) local XML files on that system.
>
> That argument makes no sense.

I love how you quote a tiny part of my argument and then pretend that is
the whole argument.

> If you only use a single machine, and a single xml file, then saving
> stuff in the xml file has exactly the same semantics as using the
> system config.

Yes.

> Normal users wouldn't use multiple xml files anyway, unless they have
> multiple different things they track - in which case they may well
> have multiple different configurations too. Imagine if your wife did
> scuba diving too, and you tracked her dives. The system config file
> actually contains things that are *personal*, like the tech dive
> settings (which may not make sense for another diver).

Yes, I agree, there are indeed possibly PER FILE settings.

> So using the system config has no upsides. Not for "normal" users. And
> certainly not for me.

But you are completely missing the point that I made elsewhere in my
email. It is entirely viable that a user has multiple divecomputers and
wants to track them in separate files. So now she has to redo her
settings for each file?

How about she wants to open a file from her dive buddy. That moron might
prefer to see dives with pO2 values and with depth in ft - but she
almost certainly wants to see what she is familiar with. Sane SI units
and no silly other graphs that provide no information to her.

So yes. The 'my way is right and everyone else is an idiot' vision of
how people want to use this software that you present above where users
fall into two easy to handle categories may be correct. But in my
personal 'my way is right and I respectfully disagree with your
oversimplified view' version of reality there may be other people out
there that have different use cases. And I intend to write software that
works for them, too.

BTW: your argument is remarkably similar to the one made by the Gnome 3
designers. "Use it as we say you want to use it, or you are not our
target audience".

> And it's confusing as hell. There are various hidden config options
> that you can't easily see.

There are? Interesting. The only confusion that I can see right now is
that early data files have no divecomputer. Then a divecomputer without
DeviceId. And then the current file format. But those are the growing
pains of software that keeps being developed.

> IMNSHO, the *only* option that makes sense as a "system option" is the
> "default_filename" thing. Everything else is simply wrong, because
>
>  (a) you want to move your settings around with your dives

I don't think that is correct. Settings and data files are very
different issues. I sure as hell don't want your settings when looking
at your dives. Feet and psi? Surface color for the deco ceiling? No
thank you. Just as you don't want Jan's settings (you made it crystal
clear that baby poop is not your favorite color).

>  (b) they are personal issues anyway, and if you track multiple people
> it doesn't even necessarily make sense to have the same options (like
> the deco GFlow/high things, which are really about personal
> preferences, I'd argue)

The preference of the user running the program. Right. 

> I realize that we did it the way we did for historical reasons. But
> they are historical, and they are wrong.

Yes, the current implementation is not optimal and has several
issues. I am not going to replace it with something that I consider
shortsighted and ignorant.

/D