For interactive use disk I/O is fast compared to the time waiting for 
keystrokes.

Only for importing and possibly report generation would disk I/O be a 
bottleneck.

Currently the file size slows down startup because the entire file is read and 
interpreted into memory.
If the file were simply a memory image (not that I’m suggesting it should be) 
it could be read in one gulp, which would be barely noticed.

In a true database implementation (which I understand is the direction it is 
moving) only that part needed to re-build any currently open tabs would need to 
be read and converted into the internal data structure, which, I suspect, would 
also be so fast as to be barely noticed, and show negligible dependence on the 
file size.


> On Jun 17, 2017, at 4:35 PM, AC <[email protected]> wrote:
> 
> Well ok, it makes many insert/update/appends and then the transaction
> causes a pile of writes.  The disk I/O is still the bottleneck for most
> applications.
> 

_______________________________________________
gnucash-user mailing list
[email protected]
https://lists.gnucash.org/mailman/listinfo/gnucash-user
-----
Please remember to CC this list on all your replies.
You can do this by using Reply-To-List or Reply-All.

Reply via email to