Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Guy,
I am not saying it would be easy. It would require effort and good
programming, but after having spent a decade using UNIX run systems
controlling multiple devices simultaneously on synchrotron based
x-ray microscopes, I am skeptical that a MacOS based G5 platform
would be that difficult to implement.
I have experienced the mouse click problem you refer to, but when
running the x-ray scopes under Unix, I was able to open additional
windows and carry out analysis near simultaneously, as well as,
effectively run additional background programs without any crashes,
something I cannot say for NT or Windows 2000, an OS I rather like
even if it is not OS X.
I could be wrong; I am not a programmer except when there is no
alternative. However, I cannot help thinking that a far more reliable
and secure system could be built around MacOS X. BeOS might have been
an even better answer if they had stayed in business, but I think OS
X and G5 processors should do just fine.
Mario
>Search the CONFOCAL archive at
>http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
>Mario is missing a significant point here.
>
>I'll not join in any argument about speeds of various systems
>except to say that in the end there's no point in having fancy
>options in your hardware unless your code supports them. Dual
>processors are irrelevant unless your code uses both, built-in
>graphics libraries mean nothing unless your software calls them,
>and so on.
>
>But driving a confocal is about far more than speed of processing
>operations or amount of addressable RAM. It depends on real-time
>access to instruments, and traditionally Unix has been very poor
>at this since it seeks to control all timing and scheduling issues
>itself. This becomes a nightmare, for example, if one is seeking
>to drive an AOTF under software control to carry out patterned
>irradiation. This also explains the 'annoying' feature that many
>Windows machines will not respond to mouse clicks etc during image
>acquisition.
>
>I'm not saying that there are no ways around it - I am saying that
>it is a problem and indeed is the reason most confocals use Windows.
>(And why the venerable Leica machines of the late '80s and early 90's
>used OS9, an optimised real-time operating system).
>
> Guy
>
>
>Quoting "Mario M. Moronne" <[log in to unmask]>:
>
>> Search the CONFOCAL archive at
>> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>>
>> Steve,
>>
>> I do all my offline processing with my dual processor G4 (and G5 in
>> the near future). Yes, it would be a lot of work to create a Mac
>> driven system but once accomplished there would be very little in the
>> way of having to account for hardware changes and system software
>> modification. It is one advantage of having the software and hardware
>> company all in one.
>>
>> If there are others out there that are interested including
>> cooperative manufacturers and software people (I know of at least
>> three regarding the latter), I would be very happy to help put a
>> proposal together for NIH or NSF to create a MacOS driven confocal
>> (not that they would fund it). How about having nearly 8 Gbytes of
>> memory for collecting data and writing it out in the background or
>> performing simultaneous deconvolution (dual processor G5 4 Gigs per
>> processor). OS X is a UNIX, after all.
>>
>> My guess is that it would cost no more than a commercial MP confocal
>> to write the software, maybe half of that. What I find truly
>> ridiculous is that the PC-windows machines that are often required to
>> effectively operate confocals or dedicated deconvolution systems can
>> cost easily twice as much as top end Mac.
>>
>> My old dual processor 450 MHz G4 that I bought for $2400 ran
>> "measured PSF" deconvolution routines 40% faster than a 1.8 GHz P4.
>> When I priced out a dual processor 2.8 GHz high quality PC, which
>> would probably still run significantly slower than a now out of date
>> dual G4 1.4 GHz Mac, it priced out close to double the Mac's. True,
> > not many people are going to care about a few thousand dollars when
>> buying a confocal, but ease of use, speed as in FPU requirements, and
>> simple maintenance are still matters of importance to say nothing of
>> memory constraints.
>>
>> I am now ending my rant. Contact me offline if anyone is interested
>> in perhaps forming a lobbying group, putting an instrument grant
>> proposal, or otherwise getting the manufacturer's to stop complaining
>> about how much it will cost to implement their software on the Mac,
>> especially the people who have some Unix code laying around.
>> Actually, that doesn't matter either; serially controllers will work
>> just fine.
>>
>> Mario
>>
>>
>>
>> >Search the CONFOCAL archive at
>> >http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>> >
>> >A couple of questions for a Friday afternoon (in Wisconsin at least)..
>> >
>> >1. How many list members would like to see a Mac platform for
>> >their confocal microscope?
>> >
>> >2. How difficult would this be to implement on an existing
>> >microscope, Zeiss, BioRad, Leica etc....???
>> >
>> >Steve
>>
>> --
> > ________________________________________________________________
>
>--
>Associate Professor Guy Cox
>Electron Microscope Unit, F09
>University of Sydney NSW 2006
>+61 2 9351 3176
>
>Until 25th July:
>Instituto Gulbenkian de Ciencia, PT-2780-156, Oeiras, Portugal.
>+351 21 446 4638 (office) or +351 91 401 5726 (mobile)
>Fax: +351 21 440 7970
--
_________________________________________________________________
Mario M. Moronne, Ph.D.
NanoMed Technologies LLC
President and CTO
[log in to unmask]
[log in to unmask]
|
