Posted by
mahogny on
URL: http://confocal-microscopy-list.275.s1.nabble.com/Micro-Manager-at-the-ASCB-meeting-tp1627182p1633781.html
Nenad has given a good response but I can't help but responding too.
as a developer of open source imaging software (www.endrov.net) my bias
is clear. but I don't develop it out of monetary reasons, what existed
before simply sucked. I hear that commercial software is tested but I am
yet to use something that is even a nice try. For example, ImproVision
has consistently ignored our bug reports(!) about OpenLab to the point
that we ended the license and are now phasing it out. On the contrary, I
claim open source is by far more tested. our community reuse a lot of
existing code that is used in several major open source programs. how
can commercial companies that all too commonly write things from scratch
compare to this?
Support - it is true that open source often does not come with a nice
manual but that forces programs to be self-explaining, FAQS and Wikis
grow, and you can commonly talk to the developers directly on IRC or by
mail. if this still isn't enough you can always fix it yourself or hire
*anyone* to do it. if you are willing to pay or have the competence you
do not have to rely on anyone. if the software is commercial and the
vendor disappears or loses interest, you are screwed. if the grants for
the developers disappear you can take it over yourself, spread the load
to multiple users or just pay up - but there is a way forward.
I do not think public funding is a problem either way. software is
write-once, use as many times as you want. the really important part of
micro-manager is the drivers and those will not need to change (the
hardware does not). this reduces the workload plentiful for us 10% of
the users who now just have to do 90% of the development instead of
100%. development of new niche applications is all about activation
energy and if the trivial work is done (acquiring 5D images, rendering,
basic image processing, storage) you will be productive from the first
line of code. this opens up for new research which is not commercially
viable (yet).
on a side-note, my plan for endrov is to bring it to main stream open
source so that more volunteers can do work without funding. image
processing is done everywhere, not just in the biology lab. See e.g.
GIMP. my advice to anyone doing niche open source is to "outsource" as
much as possible, make sure your implementation strategy takes this into
account as soon as possible. use off-the-shelf code or if you produce
nice code, try to make more people use it (release module as separate
project).
open source reduces commercial opportunity - this is a good sign! it
means you cannot make money out of something because it has been made
*trivial*. if you want commercial benefit, bring us the state of the art
stuff or we don't need you. that is the bottom line of open source.
commercial services are not excluded but you have to work for your pay.
/Johan
Kevin Ryan wrote:
>
> A perceived conflict between Open Source/free software and commercial
> products has been around for quite a long time.
>
>
>
>
> Twenty years ago when I started developing imaging software I had an
> NIH customer ask “Why can’t you _/give/_ your software away like NIH
> Image?!?” (I told him I would gladly do so when the Federal government
> started paying my salary.) Certain people in the commercial side of
> things at the time were quite upset that the government was supporting
> someone to compete with us.
>
>
>
> As it turned out, however, some of our best customers were previous
> users of these freeware packages. Best, as they had the experience
> with imaging software to recognize what it could and couldn’t do, and
> the knowledge to move up to a commercial package with
> wider/faster/more capabilities once they hit the freeware limits. They
> had the education to know what they wanted.
>
>
>
> Open Source software/hardware has the advantage of being ‘free’, but
> that freedom may cost a considerable investment in people; the grad
> student who maintains it, the technician who knows what wire does
> what, the involvement in keeping up with a wide group of contributors
> who are fixing bugs and adding capabilities. It’s a real tradeoff
> between dollars spent and the time invested to use/maintain it. I have
> yet to see a free software product whose documentation matches the
> average level of commercial software. And if the support community
> doesn’t pass a certain critical level, like Linux or the Apache web
> server, having a principal developer move on may mean the end of that
> package. Open Source/free software also tends to have a more limited
> feature set, as it tends to be driven by the developers interests
> rather than market pressure and dollars.
>
>
>
> Commercial software may not do exactly what you want; it’s a truism in
> commercial development that 10% of the customers require 90% of the
> development time, and open software may represent the clearest way to
> get some experiment specific item functioning. But for the other 90%
> of the features and customers, you get code developed over many years,
> QA tested before release, and used by a _/lot/_ of other people. There
> are often more/broader features in the commercial products, due to the
> larger number of customers requesting them and more man-years in
> development. We have something on the order of 50,000+ customers over
> the years between the various packages for our company alone, and if a
> issue or new technique pops up we hear about it and have a great deal
> of motivation to work on it. Many of the commercial products
> themselves have interfaces for users to add new capabilities, just as
> with the open software.
>
>
>
> So both free and commercial software have their niche – open/free
> software for people with more time than money, or needs simply not
> addressed by current commercial products, or who enjoy developing
> their own tools; commercial software for a wider range of people using
> more established techniques, or who don’t want to invest or even have
> the human resources for a less supported product.
>
>
>
> I do have my concerns about sponsored ‘free’ software – Micro-Manager
> (and NIH Image/ImageJ before it) are supported by NIH grants instead
> of customers or donated developer time. They aren’t developed with
> users donated time like Linux/Apache. What happens to sponsored
> software users if the grants run out? Will the support be there over
> the long haul? On the other hand, what effect on the commercial market
> and its products does government sponsored software have, other than
> reducing commercial opportunity and investment? But that’s really a
> discussion between commercial interests and sponsoring agencies,
> rather than individual products and projects.
>
>
>
> Kevin Ryan
> Senior Project Manager
> Media Cybernetics, Inc.
>
>
>
> *From:* Confocal Microscopy List
> [mailto:
[hidden email]] *On Behalf Of *Nico Stuurman
> *Sent:* Sunday, December 07, 2008 7:11 PM
> *To:*
[hidden email]
> *Subject:* Micro-Manager at the ASCB meeting
>
>
>
> If you are attending the upcoming ASCB meeting in San Francisco,
> December 13-17 (
http://www.ascb.org/meetings/), please come see us in
> booth 2009 in the Exhibition space.
>
>
>
> Micro-Manager (
http://micro-manager.org <
http://micro-manager.org/>)
> is Open Source software for microscope image acquisition. ….
>
>
>
--
--
------------------------------------------------
Johan Henriksson
MSc Engineering
PhD student, Karolinska Institutet
http://mahogny.areta.org http://www.endrov.net