Journal of Open Source Medical Computing First Call

The Journal of Open Source Medical Computing (JOSMC) is open and issuing its first call for papers. The Journal was started after the success of Linux Medical News indicated the need for a more scholarly publication. The Journal ‘…is an electronic forum for disseminating information on free
and open source medical computing.
Scholarly work on any aspect of free
and open source medical computing will be considered for
peer-reviewed publication…’
Read the editorial guidelines for information on submitting articles, editorials or features. There is a list of possible subject matter here.

Free and Open Source Basics

While relaxing this past Thanksgiving with my family, I asked my Ph.D brother how he liked Linux Medical News newsletter he replied that he liked the letter, particularly when my Chihuahua Cindy speaks, but the articles ‘are like a new language. I still don’t quite get what open source means.’ He’s not alone. It is time to review the basics.

Software is the stuff that drives computers, much like music on a Compact Disk drives a stereo, or a TV set requires a broadcast in order to be meaningful. Software is composed of odd little sonnets to a computer that are frequently laboriously constructed by a programmer. Here’s a short example of the software that drives part of LinuxMedicalNews. You don’t have to understand any of it, just know that it exists:


<dtml-in “objectItems([‘Poll Question’])”&#62

<dtml-var question&#62<br&#62

<dtml-in “objectItems([‘Poll Response’])”&#62

<input type=radio name=”responses:list” value=”<dtml-var questionId&#62<dtml-var id&#62″&#62

<dtml-var response&#62

</dtml-in&#62

</dtml-in&#62

The above is referred to as ‘source code’. Applying a computer to the source code makes the computer perform useful work. In the above case it is a small part of what makes the little poll box on the front page of LinuxMedNews work. In the terms of computer programming, it is the ‘source’ for achieving that work. As you can see from the source code example above, the intricacies of the source code can be quite complex and hence, time consuming to make. But once created, it can easily be transmitted by floppy disk, CD-ROM or the Internet and be reused by others. It can also be extended, changed or fixed if it has an error. It becomes a concentrated piece of knowledge, and one less thing for a programmer or software engineer to spend time on. Good source code can greatly accelerate what an engineer can create in a reasonable period of time.

LinuxMedNews frequently mentions the term ‘open source’. We can define open source now: it is source code that is distributed and available for a programmer or software engineer to reuse, extend or fix if it has errors. Precisely what I have done above. I’ve given you, the reader, my source code so that if you wanted to build a site like LinuxMedNews that displayed polls you could use the above programming to do it. This would save you an hour and possibly much more of work reproducing what has already been done.

Let’s pause for a moment to consider another term you may hear frequently: ‘Free’ software. I put it in quotes because it means freedom from closed-source restrictions, not free in the monetary sense. A common example is that free software is ‘free as in speech, not as in beer.’ Read more about what ‘Free’ software is here. The term ‘Free’ pre-dates ‘open source’ software and the two are quite similar except for some licensing differences which you can read about here. Let’s return to our source code discussion.

What happened with the above source code example is a demonstration of open source in action. I needed to run polls on LinuxMedNews. Howard Shaw, a colleague of mine who runs Houston Linux Users Group thought he could also use polls for his HLUG site. He wrote the source code above to fix an appearance problem in a previously written poll program and since I had a similar need, Howard sent it to me by e-mail to use on my site. He saved me a lot of work. In the future, I’ll return the favor.

‘I’ll return the favor’ is not just a haphazard statement in free software. It is the essence of free software. If you make changes to source code you’ve received, and distribute or publish those changes, you have to give anyone who uses it the same rights you have. Note that private changes for your own use remain private (see discussion below). In fact, the various free licenses require that you do so, or you don’t use the software.

The converse of this is ‘closed source’. Howard could have made things hard for me by denying me access to his source code. He could instead have said I’ll create your polls for you on my site and you can just link to the poll on my site. This would have made things quite inconvenient for me, as well as making me dependent on Howard for polls.

Closed source software is what most commercial software companies do. They usually do not give the customer source code, which is the means to produce the commercial programs. They only give the binary translations of these programs which a computer is happy with, but which is more or less useless for a person to read.

Without the source code, the customer isn’t free to fix, extend or change the software, they must rely on the company to do it for them. Whether customer changes are a priority for the company is subject to the goals and available resources of the company. There is no guarantee that the company will ‘do the right thing’ from the customers viewpoint. They do after all, have to stay in business. However, removing the restrictions on a customer fixing or extending the software, with the requirement that the fixes or extensions go back to the company, frees the customer from having to wait or otherwise be inconvenienced by closed source. The company benefits by reduced maintenance costs and free extensions for its products. More ways a company benefits can be found here.

These issues are particularly important for medical software. The same can be said for industries that need high availability and adaptability. The amount of engineering resources required to create and maintain medical software is large and expensive. Medical software has to be able to react to the ever changing winds of medical knowledge and delivery while still being reliable and cost effective. It needs to be interoperable and extendible so that novel applications such as patient simulators can be built from a stable base. These are goals that no single medical software company has been able to deliver. The resources are simply not available at a reasonable cost. Only free and open software holds the possibility of achieving these goals. That is why free and open source software is so compelling in medicine.

MedScape: TelePsychiatry

It seems like Tele-blank in medicine is all the rage. The latest is Tele-Psychiatry which appeared recently on Medscape (free login required). ‘Hi-tech video conferencing may be a cheap, effective means of providing
psychiatric care to patients in rural settings, according to a new study.

“A lot of these people would not have been seen because they don’t have resources locally or they wouldn’t otherwise travel,”…Dr. Hilty worried that HMOs might be
tempted to use the technology to replace the “real thing.” There may be economic advantages, “but the service would probably
not be equivalent,”…’

Best of LMN: Beyond This Horizon

This article originally appeared August 6, 2000 and is a ‘Best of LinuxMedNews’. It attempts to paint a picture of what a free and open source medical software future would look like and generated a fair amount of discussion. Read it in its entirety here. The amount of activity in open source medical software has stepped up considerably over the last months. Several projects, among them FreePM, GNUmed, Freemed and TK_familypractice are moving forward at a good pace. Many are at the alpha stage, with planned beta’s in 2001. What is all this activity for, one may ask, particularly if new to open source medical software. Where is it all going? What is the problem to be solved? A vision of what is beyond the sunset for open source medical software is in order.

Will Vendors of Medical Software Taste Forbidden Fruit?

Happy Thanksgiving, the following is a ‘best of LinuxMedNews’ article that first appeared 4/21/2000 My how things have changed, there are now 43+ known open source medical projects. Read the entire original article here. With 16 known open source medical software projects underway, the likelihood of eventual success for at least one of these projects in the coming years seems secure. Current vendors of medical software may be, and should be asking: Why should I go into open-source medical software? Here’s 6 reasons.

Sleep Crucial to Learning Retention

Yahoo daily news is reporting about two new studies in the December issue of Nature Neuroscience that show ‘…a good night’s sleep helps people retain some of what they learned that day…’ Which may sound obvious, but what the study also found was not: ‘…Sleep loss did not appear to make study participants forget what they learned…Instead, they seemed to miss out on a window of time in which sleep helps the brain mold information into lasting memories…people cannot make up for that critical night of sleep by
sleeping late on other days…Sleep…helps people “deal with information overload…’
All of which is a scathing indictment of current medical education and practice in which all-day-all-night-and-all-dayers are common.

WSJ: Welcome to the Wired Hospital

The November 13th print edition of the Wall Street Journal had a full page article on page R23 about Beth Israel and Deaconess hospitals experience with its electronic medical record (EMR). These hospitals have a long (1977) history of computerization and are described as forward thinking in its use of EMR’s. Editor: Unfortunately, this forward-thinking hospital apparently misses the boat with regard to making this kind of software available as free or open source. Such a closed system makes widespread adoption by other hospitals of these life-saving technologies unlikely.

The article is of interest for implementation details such as how access is controlled: by entering id, password and a security code taken from a hand-held device called SecurID which displays a code that changes approximately every minute. Dr. Slack, who is ‘among the pioneers in the field’ is quoted as saying ‘Blaming the already beleaguered doctor for being too old or too computer-phobic or too conservative is a handy excuse for bad computing… he goes on to say: ‘…In many regards, we as physicians have relied too much on memory.’ and the system is proported to reduce that reliance with the ready availability of clinical information along with the EMR. Finances and effect on the bottom line are discussed. The article estimates that the system would take $50 million dollars and several years to be built from scratch at other hospitals and reports efforts to market these technologies.

Stallman Replies to VistA Licensing Issues

No less than the founder of the Free Software Foundation Richard Stallman shares his views on the possible license controversy between the public domain VistA medical source code project and the Gnu General Public License (GPL) which LinuxMedNews recently reported on. The heart of the issue was what happens when the completely un-licensed VistA source code comes into contact with GPL’ed source code.

“I would like to respond to the main question in the article and also
correct a couple of mistakes in the article.

I am not a lawyer, but I have spoken extensively with lawyers about
copyright questions. Presuming that the VistA software is in the
public domain, if you combine it with a GPL-covered program you must
release the combination *as a whole* under the GPL. Using the VistA
code in this way is allowed because public domain status permits
practically anything.

However, the specific code that was in the public domain remains in
the public domain. In other words, the fact that person A released
the VistA code in a GPL-covered combination does not stop person B
from using the VistA code in some other way.

You could even extract the VistA code from the GPL-covered combination
and use it as public domain material, as long as it really is the
unmodified VistA code. If you want to use the VistA code as public
domain material, the safest way is to get a copy of the original VistA
code, because if people have changed that code since, they don’t have
to put their changes in the public domain. Still, in principle, the
VistA code remains in the public domain even inside the GPL-covered
combination.

Another way of putting this is that the GPL is not “contagious”. The
GPL applies through inclusion, not through contact. It applies to the
combination because the combination (given the assumed scenario)
includes some code that was released under the GPL. But the VistA
code retains its own status, despite being in a combination with the
GPL-covered code.

The same conclusions would apply if, instead of the public domain
VistA code, we were talking about a module released under the X11
license.

There are other licenses that have different restrictions,
particularly with regard to commercial use of software such as the
FreeBSD License. The Free Software Foundation does not consider these
licenses to be ‘Free’ licenses.

Actually we do consider them free licenses. Both the original BSD
license, and the revised one preferred by the FreeBSD developers (and
adopted by Berkeley a couple of years ago) qualify as free software
licenses, like the X11 license. We have used code available under
these licenses as part of the GNU system since the 1980s.

These licenses are non-copyleft free software licenses. See
http://www.gnu.org/philosophy/categories.html for more information
about this and other categories of software, and see
http://www.gnu.org/philosophy/license-list.html for more information
about the status of specific licenses.

Please note that there is no GNU Public License; you probably had in
mind the GNU General Public License, or GNU GPL. It is common to drop
the “GNU” and write just “GPL” when there is no danger of confusion.”

Examples of Health IT Failures

Two new subject headings has been added to LinuxMedNews: Health IT Successes and Health IT Failures. It is hoped that the first category grows faster than the second, but this compendium of failures by medical informaticist Scot Silverstein shows that this category will be alive and well for some time to come (Can you say 43 million dollars up in smoke?). His site also offers some perspectives on the problems of Health IT.