[Libreoffice-bugs] [Bug 55960] Filling cells with incremental numbers – wrong numbers

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=55960

--- Comment #24 from b. <newbie-02 at gmx.de> ---
(In reply to Mike Kaganski from comment #22)

thanks for your fast reply, 

> I doubt that this is "info about the significant decimals keyed in by the
> user". Rather, it's the number rounded to some number of decimals on save,
> which happens to coincide. Or even possibly a result of smart
> float-to-string algorithm finding shortest string that round-trips to the
> same float.

whatever it is - you'll know better than me - it's a valid starting point for
the proposed algorithm, 

> if you
> can suggest an optimization that gives better precision by re-arranging
> calculations, without performance loss - it's fine to discuss in a dedicated
> issue like "Function FOO can give better precision", with analysis of
> correctness of suggested optimization, and comparison of performance and
> precision of old and new ways.

this is too high a demand on me, since i am neither c-coder nor performance
specialist nor LO developer, 

(besides i tried to give 'my best two cents' in the previous post,) 

an important basic question is whether you are allowed to comment on problems
if you cannot do it better yourself, Adolf Hitler said 'NO!', i say 'yes', 

["Only he who can solve a task better is entitled to criticism". - Adolf Hitler
Speech on May 1, 1934 in Berlin, in: Max Domarus (Ed.): Reden und
Proklamationen, 1932-1945, Vol. 1, p. 379

Source:
https://beruhmte-zitate.de/zitate/128459-adolf-hitler-nur-der-ist-zur-kritik-berechtigt-der-eine-aufgab/]

so, if you can accept that even people who are 'idiots' in certain areas are
allowed to bring up strange results from these areas and may make - even
idiosyncratic - suggestions for solutions, among other things because mankind
has mainly developed in this way ... and because you can sometimes free
'specialist idiots' from their narrow-mindedness ... then we can continue
talking, 

in the sense of goals and correctness - imho LO / calc should! have goals how
to solve calculations, and theese two: 'math-correctness' and
'ex$el-compatibility' i encountered in discussions here again and again,  

'performance' is a dilemma, for 'small tasks' like filling 100 lines with
discrete values the 5 nanoseconds? extra computing time needed for step-by-step
rounding is certainly the smaller the problem, in contrast to the hours and
days we spend explaining the 'necessary errors' to users, in large sheets with
thousands of values and calculations the computing speed becomes of course more
important, but there the loss due to the immensely complex troubleshooting also
increases and the consequences of undiscovered errors can be terribly harmful, 

[in which city do you live? LO is in use at italian military? consider they
decide to bomb greece out of europe to become 'the only one largest problem in
europe', they use calc for the target calculation and nuke out your city ...]

it is not my choice or decision how LO and calc and 'the developers' deal with
these problems, but my urgent recommendation to try to produce satisfied users
with correct results and to try to get rid of fp-rounding-errors and to get rid
of the miserable compatibility to ex$els errors ... while keeping doors open
for users to migrate ... towards LO ... and for that 'more accurate results
than ex$el' would be a great invitation ...

P.S. 

tdf#129606: correct, well done, keep going ... 

P.S. II 

i'd suggest to think about general solutions rather than individual bugs ...

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20200903/5e1dfbf0/attachment.htm>