On Thu, 17 Mar 2011 22:54:13 +0100
Pavel Sanda <sa...@lyx.org> wrote:
> Michal, this is not the main issue. let me explain this in more detail
> so you dont get angry about my silence.
>
> many of us want to discontinue the schizophrenia of two different
> windows installers and i'm trying to manage affairs so that we finish
> with single one.
... but why are you telling me this after I've put (wasted?) so much
time in building and testing this installer? It's not that I was hiding
or anything... knowing this beforehand, I'd have planned things
differently.
> Uwe's (alt installer maintainer) decision whether he is able to
> continue with the current Joost code will be vital for the final
> result, but for this moment i would like that all development/testing
> energy goes into the (yet) unreleased official installer. it should
> already support new dictionaries, it should demand less windows
> privileges, the code should be simpler than both previous 1.6
> installers and so on.
I don't know the details of the new installer (well... almost no one
does, as it does not exist yet), but I have my reasons for always using
the AltInstaller (being forced to reinstall the whole system after using
the 'official' installer to update LyX is among these).
These are just my personal reasons, not really that important,
but my point is that if you want to convince users to use the new
installer, just try it with providing the features they want first(!),
and advertise these well. Blocking out the paths you don't prefer
instead, is not a good idea IMHO - especially when this tactic leads to
offering no path (installer) whatsoever.
> i was accepting the last beta4 contrib binaries as a complete
> emergency case and want that users go with official installer for rc1
> even for the price of 2 weeks delay...
It's not that easy for the ones paying this price, you know...
being on the receiving end was in fact the main reason I got involved in
creating my own installer - beta3 was just too buggy to use (and even
to test), and no one really cared about the casual users in this regard.
Now, after reading your mail, I still don't understand what would be
wrong with providing AltInstaller now and add other installers later.
Also, as the new installer is completely untested (because none has
been built, even as of today), how is that fair to only test it in the
'release candidate' stage, where everyone is expecting something stable
enough to work?
I mean: the version with completely untested installation process
shouldn't be called 'release candidate' at all. According to Wikipedia,
in release candidate '... all product features have been designed, coded
and TESTED' (emphasis mine).
Please, rethink this all - I hope this mail will help somehow...
otherwise you are going to needlessly alienate your users, and one day,
some EMAltInstaller ('Even More Alternate Installer'), developed abroad,
will be an added bonus :)
> if you have already set up compilation environment and have some
> energy i would propose that you experiment with the new installer,
> there will be many new bugs i guess, just read the next mail in
> thread...
Well, after my work being thrown away and all my comments about
vs2010 being largely ignored (I mean by the ones responsible for such
a huge change, such late in a release process), I doubt I will have any
energy left - I am really busy right now, and not even knowing my work
will have any sense isn't going to help...
I'll try to provide next versions for AltInstaller via hotfile, as
long as it still works or until Uwe comes back, that's for sure. As for
the other stuff, I'm only saing 'maybe'. LyX is a moving target right
now, so I'd rather wait with my experiments, until at least README.* and
INSTALL.* files stabilize...
Regards,
Michał Skrzypek
