OSTC firmware update prompt and planning for 4.4

[Dirk Hohndel]

On Mon, Dec 29, 2014 at 08:08:47PM +0100, Davide DB wrote:
> On Mon, Dec 29, 2014 at 6:19 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
> >
> > What's a "Real SAC graph" and who wrote it was "nearly there"?
> > I must be running low on coffee (definitely running low on sleep), but I
> > have no recollection of this.
> 
> Sorry I thought one thing and I wrote a completely different one.
> I was speaking about real time average depth graph/value: instead of
> drawing a meaningless straight average depth line, it would be useful
> having a line who follow the average depth calculated at every sample.
> I could find again who wrote that it was implemented and it (just) needed a UI.

Ahh, OK, this I remember.

> > It may not work for the way you dive and the way you want to record your
> > data. But trust me, it works perfectly fine for a LOT of people. I have
> > quite a few multi tank dives in my dive log with many switches back and
> > forth, done with multiple computers and all imported with Subsurface.
> >
> > So please don't make such generic statements.
> 
> Dirk, please do not misunderstand me. have you seen how much
> attachments I added on the track ticket? I spent one day making test
> and writing the ticket so it's not a generic statement.

The way you wrote it in your email (quote "multiple DC on tech dives
(multiple cylinders and gas switches) does not work" end quote) is simply
not correct. And that is what I responded to.

I appreciate the detailed explanations in the ticket and once I am done
with the other two hundred fifty seven items on my todo list I will
investigate in detail what is going wrong for you there.

In the mean time I think by FIRST importing and THEN modifying things
should work for you as well. I could be wrong, though.

> Nevertheless I had a lot of fun doing it :)
> I do not pretend anything of course.
> I tried with different devices, all used as bottom timers, and I
> always get the same result.

But you always did the same thing, I believe. You always import from one
dive computer. Then edit things. Then import from the next dive computer.
And that appears to create a scenario that confuses our algorithms -
mostly because our algorithms assume that the different dive computers
were on the same dive and therefoew give you reasonably consistent data.

Since you use only bottom timers which in general end up having incorrect
gas information (and also use a pSCR which is ill supported at the
moment), the combination of all these different variations from what Linus
and I tended to use and do (which is how the code was designed and
written) clearly triggers some corner case we didn't consider.

> Bottom line is that I have to import all my devices before starting
> entering data.
> You could just close that ticket simply explicitly stating this in the
> user manual.

I could, but that would mean that I discard your input. And that I don't
want to do. Your input is important and your use case is clearly real (as
indicated by the fact that that is what you do when diving). So I need to
find the time to figure this out and fix it. Not sure if this will happen
before 4.4.

> > It is possible that Subsurface could make this easier, but I have not seen
> > a design proposal of how that would work, precisely. And whatever that
> > proposal may be, it needs to make certain that it does NOT break the
> > currently working code for people who dive with multiple actual dive
> > computers that are correctly set up for the gases used on the dive, etc.
> 
> I did not think to write a proposal. I was waiting to be able to reproduce this.
> I will risk the firing squad later with my proposal :)

There are very few firing squads involved in Subsurface.

> >> After 4.4 it would be nice having a rehash of  stats tab. It's strictly
> >> linked with the multifilter feature. I can draw some sketch, if you like.
> >
> > Your sketches have been extremely helpful in getting us a very nice
> > multifilter solution, so by all means, yes, please do so.
> >
> > I know that Tomaz has done a lot of work on the statistics that are locked
> > away on a computer he currently doesn't have access to :-(
> 
> We will wait for Tomaz implementation :)

He told me just this morning that he may have access to that laptop again
soon... we'll see.

/D