Archiving old transations

[Jonathan Kamens]

On 06/21/2013 03:54 AM, Colin Law wrote:
> How long does it take to load the file with no tabs open?  Time for
> how long it shows "loading user data" on the startup screen (with no
> tabs open).
It appears that if I had not already archived many years of transactions 
because of the start-up speed issue, it would take ~10 seconds to load 
the user data every time I started the app. And that's just loading the 
user data -- it's around another eight seconds to load all the reports 
and do the other non-data initialization.

By way of comparison, LibreOffice Writer takes four seconds from when I 
launch it until when I have a window ready for me to type in. Firefox 
takes even less time than that. GIMP takes less than eight seconds when 
I start it up cold, and less than four seconds when I close it and 
immediately open it again. Picasa, which is managing far more data than 
GnuCash is for me, starts up in 12 seconds total, and that's including a 
speed hit from running inside VirtualBox.

If supporting archiving old transactions is such an albatross, then I 
suppose another option for improving startup time would be to figure out 
how to make it take less time to do the non-user-data initialization. 
For example, I suspect that GIMP does some sort of caching of plugin 
initialization so that plugins can be initialized faster after the first 
time, and that's why it starts up so much more quickly the second time 
than the first.
> Is that 21 MB of compressed xml?
No, uncompressed, because I transfer it regularly between machines with 
rsync, whose diff algorithm works much better on an uncompressed file. 
Loading a compressed file takes about the same time as an uncompressed 
file; as far as I can tell, it's the processing, not the disk I/O, that 
causes most of the delay.
> They are under
> no obligation to do anything for anyone.
I am not talking about obligation. As I've already said, I assume that 
most people who maintain software actually care about giving their users 
the features they want and need, and I believe that this feature falls 
into that category.

Of course, there are exceptions to that rule. The maintainters of GIMP 
have written off a (large) class of users and have made it clear and 
explicit on numerous occasions that they have no intention of putting 
any effort into making their app more hospitable to those users. I think 
that's an unwise thing for them to do, but it's their right, and at 
least they're up-front about it.

That's not what has happened here. No one has ever said, as far as I can 
tell, "We don't think the kind of users who care about GnuCash taking 
too long to start up is our target demographic, so we're not going to do 
anything to help them." Rather, what has been said is, "Yeah, that's a 
problem, and we're working on a fix" (and then no fix has arrived), or, 
"Yeah, that's a problem, but no one is actually working on fixing it," 
or, "Shut up and stop complaining" (mostly by non-developers).
> Telling them that they are
> rude and patronising is not likely to encourage them to help you.
> More likely the contrary in fact.
I was polite about it for many, many years. And as you just pointed out, 
you're not a GnuCash developer, so apparently I /didn't/ tell the 
developers that they are being rude and patronizing.
> I think the reason that it has not been addressed is that most do not
> see it as a problem that needs to be fixed.
Several GnuCash maintainers over the years have acknowledged the problem 
and indicated at least a theoretical desire to solve it, so I think you 
are wrong about that, and I think that since you are not a developer, 
you should perhaps stop trying to speak for them.

There are features in my Thunderbird add-ons that "most" users do not 
use, but they benefit enough of my users that I was willing to put many 
hours of work into implementing and maintaining of them.
> I certainly do not want
> to archive old data out of my main file.  I like to have access to my
> complete history so I can easily search for old data.  I often find
> myself, for example, searching to find when I bought something, and
> love that fact that I have 15 years historical data easily available.
The fact that there are some users who do not wish to archive their old 
transactions has little to no bearing on whether there are enough users 
who do to justify making it possible.

The facts that three different developers over the years have 
implemented solutions to this problem, and that many people have asked 
about it over the years, would seem to imply that there are a 
significant number of users who want to be able to do this.

   jik