On 12/13/2012 06:32 PM, Uwe Stöhr wrote:
Am 10.12.2012 18:59, schrieb Richard Heck:
So what does my change break?
Here's my problem: If we do things as you propose, then we have
different layout files with the same
name under 2.0.5 and 2.0.6.
I understand your point. But as said, who really downgrades from a
working LyX 2.0.6 to 2.0.5?
I think this question has been answered. The worry isn't so much that
people downgrade (though they do, if a new version introduces a bug that
really bothers them). It's rather that people often use several
machines, and those machines may not have the same LyX version on them.
I keep my main machine at home very up to date, for example, but not
always my laptop or even my work computer. Let alone my wife's Mac,
which I might use on a trip. This is very normal.
We in the past always added styles to layouts because our major
releases only take place every 2 years while the document class
progress is much faster.
Please have a look of the layout file changes during the 2.0.x cycle.
It was not only me who added a style to layouts, others did that as
well and for scientific classes we will have to do it anyway for the
reasons I stated.
But as this issue is so fundamental, I request that you bring this to
the users list and decide then.
Feel free to start a discussion on the user list.
So people using two different minor versions on two different machines
will find that a single document won't even load properly on the
older version. And that passing it back and forth between the two
causes all kinds of problems. Specifically, files created with the
new ACM and europecv layouts will not load properly in 2.0.5.
- the changes for europeCV are cosmetic ones: I only added styles for
things one currently has to specify in the preamble. I can retract my
patch for this document class.
This looks to me like almost a format change. When we do this kind of
thing, we should have lyx2lyx that will find the stuff previously in the
preamble and convert it. I think Jurgen has just done this for some
things in beamer.
- for ACM we will have no other choice because your paper won't be
accepted for submission without the new styles. So if you reject my
patch for ACM, you force users to read the submission guidelines after
they notice that their submission was rejected. They will then find
out that they will have to add things as ERT. (But only experiences
users know how to handle ERT.) Is this our aim?
My proposal is to ship the new layout, but with a new, versioned name.
I'm opposed to this, see my arguments in the new thread I started.
I'll respond to that separately, though see below.
For the document class files: I state it the last time: You have to
fulfill the submission guidelines. The journals don't care of your
OS and versions. If you don't fulfill the guidelines, you won't be
accepted, point!
Yes, but you are assuming something that does not seem evident,
namely: that everyone who is using these class files is doing so
because they are planning to submit their paper to a journal that
uses that class file.
Of course, because that is what these document classes and thus our
layouts are designed for.
They may be designed for it, but, as many people have said, that just
doesn't imply that that is always what they are used for. As people have
pointed out in the APA6 thread, the APA classes get used for thesis
formats by some universities or departments. It would not surprise me if
there were similar examples involving the journal classes, especially
ones that have to do with conferences. Maybe some obscure journal uses
the ACM class, and it's sticking with the older one for now. Do we know
this is not true? Or what if I am running a conference and I've said we
want submissions using such and such a class? Am I going to change that
just because someone has decided to release an update to that class?
Maybe so, and maybe not.
As I've said, as long as the updated layout classes are available,
people can use them. I just don't see the point of breaking a bunch of
other stuff to make that very slightly easier.
That is not true. People have old files that got rejected, but they
keep them in
that format anyway. Those files should not stop working because they
upgraded LyX.
As I stated, except of the case with Springer old files keep
compilable. None of may changes break the compilation of existing
files. Also pre-prints, editor notes, etc. work fine.
Unless I'm mistaken, they do break compilation, because the LaTeX files
that we output will not compile with older versions of the class files.
Yes, I understand that, with the journal layouts, this will not be an
issue for most people who are using them, because they will have to
upgrade to the most recent version in order to submit their papers to
certain journals. But this is not the only use case.
I'll continue this in the other thread.
Richard