Where are all the TEM Applications?

If you have anything to do with the field of computer software, you know that there have been monumental changes in the accessibility to developers over the last 10 years.  When I started doing commercial programming almost 20 years ago, your choices were limited.  There were only a few compilers, they were fairly expensive, difficult to use, and they still always seemed to be missing important things that you needed.  I’m not exaggerating that there was a several month learning curve to do anything useful.

These barriers don’t exist any more, at least not in the web world.  The number of tools, languages and IDE’s, most of which are free, has led to thousands, if not millions of applications available, again mostly for free.  But there must be some kind of barrier remaining in the TEM world, because I don’t see very many new applications having come out in the approximately 5 years I’ve been “out” of TEM.  So I feel the need to ask the question:

Where are all the TEM Applications from the last 5 years?

I should first describe what I mean by a “TEM Application”.  To me a TEM application is anything that improves the efficiency or simplicity of running the instrument and/or manipulating your data from the instrument.  This doesn’t include the basic software that runs the instrument and its detectors, like Gatan’s Digital Micrograph, or EDAX’s Genesis, or FEI’s Tecnai/TIA, which I refer to as “platforms”, because they provide basic functionality (although it may be fairly high level) through manual operation.  Applications generally involve some level of automation – doing with software what you would normally do manually – and/or integration – giving access through a single interface what is otherwise done through several interfaces.

So what applications are out there?  Here’s what I found from commercial vendors that I would classify as “TEM applications” and not just functionality of their standard system:

• Gatan:  About 10 applications written for Digital Micrograph, including TEM AutoTune, Digital Montage, HoloWorks and EELS Advisor.
• Thermo Scientific:  COMPASS (Multivariate Statistical Analysis),  XPhase Compositional Analysis, Spectral Match, EDS Feature Sizing.
• Oxford Instruments:  INCAEnergy TEM phasemap, spectrum matching and Cameo+
• Olympus Soft Imaging Solutions:  8 iTEM Solutions, including strain analysis with CBED, diffraction analysis and Virology.
• FEI:  Inspect3D and Explore3D, electron tomography acquisition and reconstruction.
• JEOL:  TEMography 3-D acquisition, reconstruction and visualization software, JEOL Automated Data Acquisition System (Jadas) and JEOL Archive Management System (Jams).

Additionally, there are several applications available which were developed in the academic community:

• SerialEM – tomography tilt series acquisition from the Univ. of Colorado.
• Leginon – automated acquisition of TEM images from the Scripps Research Institute.
• NCMIR – National Center for Microscopy and Imaging Research has a number of applications for biological imaging, including a number developed at various research centers around the country.

There are also a number of small, highly-specialized (any often very old) applications available at a number of government labs:

• National Institute of Standards and Technology
• Argonne National Laboratory

Beyond that, I’ve only found a few small, usually very limited distribution apps done in individual labs.  Of course I’m sure I forgot about or missed some, so if you have or know of other significant applications, please let me know about them by posting a comment.

Many of the above apps already existed 5 years ago, and most of the new ones are centered around electron tomography.  I don’t see much new out there in other areas like spectroscopy, lattice imaging, automation/productivity, database archiving, etc.   I have a hard time believing that TEM users out there aren’t doing anything but tomography, and that they don’t need significant new software to help them with their work.

When I effectively left the field 5 years ago, there was huge promise for applications.  Microscopes were just getting stable, advanced, open software interfaces (Tecnai and FasTEM), and there were a number of equally suitable platforms for application development (Digital Micrograph, TIA and iTEM).  Based on our intentions at the time with Emispec, and based on what we heard from other manufacturers and from customers, we expected to see a flood of new application come out.

So what happened?  Well, my guess is that major hardware developments, including FEI’s Titan, JEOL’s in-column filter and FIB’s, silicon drift EDX detectors and new CCD cameras took so much development to get to the market that applications development was put on the back burner.  Unfortunately, I don’t see much effort to move applications to the front burner, but if I have anything to say about it, we’re going to cook up lots of new applications now that we’re back in the TEM game.

The forest for the trees

As you might have guessed from reading JK’s recap of the last few years, the progression through the chain of Emispec, FEI and Enoetic that has now led to AppFive has been a pretty wild ride.  I was a fairly public face of Emispec — if you received either ES Vision field service or a demo, there’s a good chance we met personally — so I should chime in as well.

As JK indicated, everyone at Emispec was very much forward-looking.  We’re looking to renew that vision in AppFive, with a focus on developing software applications.  Yes, we put great effort into designing good architectures and platforms, but what got all of us excited at the end of the day was making programs that manifested themselves in the physical world — whether that was through automation of the column and its associated detectors, analyzing experimental data to reveal new insights, or relieving users of some microscopy workflow tedium.

Times have changed since the Emispec days, and pulling back from the microscopy world for a few years to work as Enoetic was the “forest for the trees” moment that made this crystal clear.  From end-users to programmers, what people expect from software has changed.  If nothing else, the web’s collaborative-mashup culture has introduced the concept of “reuse through interaction” as opposed to “reuse through inheritance.”  That’s significant.

Now I’m not advocating a Facebook or even a web-application for EM, but what I am saying is that AppFive is excited about helping our customers:

• Break their big jobs into smaller tasks
• Picking the best tools to get those tasks done
  • The best tools might be software applications you’re already using
  • Or we can make custom software for you
• Unifying any disparate tools into a single experience

So even if you don’t need all-new, custom software — if you’ve ever been sitting in front of your computer working on a microscopy experiment or analysis and wished your software worked a little differently — we want to hear from you.

Back to Familiar Territory

It’s been about five years since most of us at AppFive have had any direct connection with the microscopy community.  We’ve been on an interesting path since the days when we felt very much in tune with the EM world as part of Emispec.  So for those of you who might have wondered where we’ve been (maybe I should say “both of you”), let me fill in some details.

Many of the days over my 10 years with Emispec were wonderful.  We got to interact with many, many interesting people from all around the world.  We got to work on challenging, interesting problems and some cutting-edge research.  We had a nice mix of software, hardware and microscopy.

But, believe me, there were difficult days, too.  We had to try to interface with dozens of older instruments with sometimes sketchy documentation.  We had to deal with countless versions of hardware and software with questionable (or just plain nonexistent) compatibility.  And we had to try to innovate while still supporting all our existing customers.

Those of you who knew us at Emispec knew that we were dreamers — idealists.  We couldn’t just incrementally improve our products and bring out an enhancement here or there.  We wanted to really move things forward, push the envelope, and make a difference.  So when the opportunity arose in late 2003 to merge with FEI Company, we saw it as a way to finally realize a lot of the dreams we had.

We thought that being attached to a column manufacturer would give us the level of access to the instrument that we needed to do the kind of control we thought was necessary.  We thought that a company with the development resources of FEI would enable us to focus on our specialty, which was writing applications.  And we thought that FEI’s huge customer base would feed us a neverending stream of interesting and relevant projects.

Well, somehow things didn’t quite work out how we expected.  I don’t want to try to explain why here, partly because I don’t completely understand it myself, and partly because it’s too complicated for this limited forum.  But I will say that for myself, part of it is that I found it difficult to work for “The Man” (a term popularized here by Steve Kim).   It was pretty obvious to myself and to several of the other ex-Emispecers that we needed to get back to our small business roots after FEI Tempe closed.

So this is when we thought we would try something new, which came to be known as Enoetic.  We laid out several requirements for the new venture:

• Software only – try to avoid past problems trying to support computer and detector hardware
• New market – for something new, try something outside of EM, and maybe outside of science entirely
• Make use of our unique skills – we thought we should do something that made use of our specialized scientific programming skills
• Try to make a difference

I can’t tell you the whole reasoning process, but we decided to create a consumer desktop product for organizing digital photos.  What would be unique about it is that it would use the content of the photos, along with all sorts of sophisticated image processing and artificial intelligence, to automatically classify your photos.  We saw all sorts of benefits of this product, the two major ones being that it would be fun and challenging for us and it would also solve a problem experienced by millions of digital camera users.  So we went off for a year to create our product, called Photology.

The only problem with our plan was that after Photology was released, it didn’t generate enough revenue for any of us to make a living off of it.  The product worked, received good reviews, and people seemed to be interested and seemed to like it.  So what happened?

I think we made two fundamental mistakes.  The first is that we didn’t see clearly enough the direction that consumer software was going, even when we started the project.  It seems obvious to us now that consumers don’t want to buy software;  they expect to get it for free (at least in the beginning).  There are several business models by which you can generate revenue from consumer software (ad-based, freeware/premium, subscription services, etc.), and they don’t involve selling the software at first release.  Not to mention that the number of hits you need to support any of these models is not in the thousands per month, it’s in the hundreds of thousands, or even millions.

But the main reason I think Photology didn’t sell is that we didn’t know our market well enough.  Part of this is reflected in the first reason above, where we didn’t know how our market gets its software, how many hits we would need, and how difficult it would be to get those hits.  I realize now that we also made assumptions about our market that turned out to questionable.  Namely, we assumed that our customers had a problem organizing their photos, and also that we would give them a way they would prefer over how they currently do their organization.  As soon as you can’t count on those assumptions, the product loses a lot of its value.

So what’s the moral of the story?  I would put it most simply as stick with what you know.  I realize that this doesn’t sound very attractive to aggressive, risk-taking entrepreneurs out there, but I think they would be more successful overall if they followed that philosophy.  I don’t want to imply that you should never attempt things you haven’t done before – how boring would that be?  But I have to stress that you should never base any significant portion of your business model or career choice on things with which you don’t have considerable experience.

So here we are back in our familiar territory.  The microscopy world is small enough that one person can get a pretty good grasp on the entire market.  Since we were thoroughly entrenched in that market for a long time (myself about 20 years), we know most of the people and companies already.

And there are a lot of good people and good companies out there.  I’m looking forward to reconnecting with them.

Is this thing on?

Welcome to the all-new blog for AppFive!  We’re really excited to have you visit and hope you become a regular here.

In case you haven’t checked out our website, I invite you to do so — it should answer a lot of questions you might have about what AppFive is all about.  In a nutshell, we’re a new company dedicated to serving the software needs of the microscopy market.  All of our members have extensive experience with electron microscopy (EM), so we’re sticking close to that right now.

We’ve met some great people throughout the EM community, and we’re thrilled to be getting back into it.  We want that excitement to carry over into this blog, so please leave any feedback you have in the comments.

With that short introduction out of the way, we’ll start rolling with the blog content…