v1.17A

[Kristian Nielsen]

> Date: Tue, 16 Nov 1999 18:43:23 +0100
> From: Farhi <farhi at ill.fr>

> I've installed the version 1.17A of McStas.

Great! I realize that I may have forgotten to warn you that this is
again an ALPHA version, meaning that it may be unstable, etc. Your
comments on problems are very useful, I will fix the problems you
mention in the upcoming v1.2 proper release.

> Problems :
> 1-Definition of components change sometimes (for instance Guide2). The

Yes, I was aware of this but it is good to be reminded! I will fix this,
probably by renaming the new Guide component to something else. For now,
just delete it or copy the one from the unofficial download page on top.

I actually try very hard to be 100% backward compatible in new versions,
so if you saw any more problems than the Guide2 component you mentioned,
please let me know.

Anyway, my plan is to implement optional parameters in components,
ie. parameters that have default values if the user does not specify
them. This would make it easier to extend components while maintaining
backwards compatibility.


> 2- Some help sections in components have disappeared since previous
> versions
> 
> Undocumented :
> Channeled_guide.comp
> DivPos_monitor.comp
[...]

Yes, the missing documentation is the main thing that I need to finish
before I can release the next version.

> 3- The MCGui seems to work ok  My comments are : would it be possible to
> make a-Highlight syntaxing, such as in nedit or vim, with C based
> template and MCSTAS macros added  By the way, it is possible for

Of course, but it would probably be a lot of work.

> instance to customize the 'Shell' menu of NEdit in order to execute some
> mcstas operations.

Yes, I think this is a better way to proceed. Anyway, there should be no
problem using another editor instead of the really stupid one in
mcgui. Mcgui will automatically recompile if the instrument has changed
on disk, independent of what is in the internal editor.

> Perhaps a good option would be to optionally use the EDITOR variable,
> and use the mcgui one as default. mcgui could then check for changes in

Yes, this is a good idea. The mcgui editor was only added because it
could be done in two lines of code with Perl/Tk, I am not sure it is
really a good idea. For serious use one would definitely use a better
editor.

> file and warn if unsaved before compiling.

> b- I guess you intend to add some copy/paste and search/replace features
> sometimes... (with Unix mouse text mode right now). Ctrl-C, X and V seem
> to work ok.

Yes, maybe.

> 4- I've tried to obtain the HMI Vitess program. No way to compile it on
> my Mac PPC G3. And it appears quite monolythic (no way to create/modify
> components or instrument structure)

Ah! I also tried it, and I agree with your impression. Apparently they
are not willing to let anyone have their source code, which is in my
opinion a very serious limitation. And also a great pity, since some of
their components seem to have received a lot of effort. Overall, VITESS
appears to be still a bit behind programs like McStas, MCLIB, and SHADOW.

> 5- We've planed a Monte-Carlo unofficial workshop on
> 
>                            January 21th 2000.
> 
> You're invited of course. The aim is to gather efforts concerning MC
> simulations (i.e. Shadow, Vitess, McStas, Restrax, ...) even if I think
> that some dev team won't be keen on cooperating.

Ok, thanks for the notice, I will make a note in my calendar. Hopefully
I can participate, it sounds quite interesting. Also, I think (=hope)
that McStas v1.2 will have been released by then.

> For reflectometers, handling neutron interferences is not straight
> forward yet (need to propagate the neutron phase also)
> For SpinEcho, we should get further in designing magnetic components...

Yes, this is an interesting direction of development, with entirely new
challenges. It is very good that effort is being spent on this at the
ILL, I am not aware of any previous major effort on spin in Monte Carlo.

> Also we need to create some efficient samples for testing instruments.
> Powder is ok, need crystals...

I am currently working with some other people at Risoe on very realistic
single-crystal samples (many reflections with individual structure
factors, multiple scattering, mosaic, ...). And similarly for
powder. Expect something good in a few months, maybe even at the ILL
Workshop (no promises though!).

> 6- After discussion, the 'chiefs' think it's better to have a specific
> ILL page for instruments, components, and documentation. Links would
> reach the Risoe site.

Ok. In any case, in the future I will be more quick to put contributed
components on the web page, especially if it just involves a link to the
ILL pages.

Thanks for the comments on the new McStas version, I will fix things
before the official release.

 - Kristian.