What I really wanted to do was to read a review on NMR software. I have been waiting for more than a decade; never seen a web review. During this prolonged period the things I wanted to read have increased up to the point that I am now able to write the reviews by myself, and much more. I will explain why NMR software is the most useful of all softwares, why nobody really cares about it, how you can use it, how you can get desperate with it, how you can write your own and why. Add your comments!
Friday, December 15, 2006
Thursday, December 14, 2006
A well written list of NMR freeware can be found at:
It's 3 years old, which is, unfortunately, too much for such a rapidly evolving field. Apart from this inconvenient, it's nearly perfect. The only (minor) defect is that the list contains more than one entry per operative system. My ideal list would just contain a unique recommendation per OS.
The questions is: why there is so much NMR freeware around, that all web lists are rapidly outdated? It is well proven that the authors of freeware never become famous, and I mean any kind of freeware, I don't mean the NMR niche. If NMR freeware is never celebrated, and if the author of any free program is never remembered, the probability of becoming famous writing it are zero. Only one word can explain why people writes it: money. They write it for the money and they stop writing it when they don't receive money anymore. This has been my case and the case of all the NMR freeware I know about. I used to receive my regular salary when writing freeware and renounced to it to write commercial software.
Unfortunately, for the users, the day arrived when the sponsors realized that this kind of software brings no fame. The only hope is to become the standard of the field. NMRPipe went near to that status, yet was limited in platforms (only UNIX) and in applications (mainly 2D). All that the sponsor can get is a publication, written by the author himself. In some cases they may add a poster or a follow-up paper, and by that time they realize they are scraping the bottom of the barrel. No external author ever cares to write a review. Even youtube movies get their reviews, but NMR softwares don't. Vomiting as it can be, this is the reality. When the sponsor realizes that even the stinkiest movie gets more attention than the best-crafted NMR program, that's the moment when the flow of money stops. And when the money arrives no more the programmer (being him no saint) quits.
I don't mean that it would have been different, but it certainly wouldn't have hurt, if the beneficiaries had cited more frequently their softwares, reviewd them, evangelized them. The beneficiaries of freeware also receive their salaries and their grants, therefore their choice was not dictated by economical motivations. If they had really loved the free software they'd a least tried to save it. They didn't love and didn't care. They decided for the death of NMR freeware and, if it was not for the residual stupidity of the sponsors, they'd had almost fulfilled their intent.
How a software dies? It is constantly fighting on four fronts:
- advancements in NMR
- changes in file formats
- changes in the operative systems, processors, compilers, libraries, etc...
- competition from other NMR programs
The lethal bullet can arrive from any of these sources, but the first cause alone would suffice. NMR is so vast a field that all programs have at least one hole: they never find the time to cover the whole gamma.
At the risk of being liable, today I uploaded the definitive version of SwaN-MR at http://www.inmr.net/index.html#DOWN. The sponsor forgot its existence. I wrote it but it's not mine. If you are so poor that you couldn't buy a new Mac in the last four years, chances are that SwaN-MR will run on your old machine. If you are so lucky to own a vintage 68K Mac (mine died), look somewhere else.
Tuesday, December 12, 2006
For the most part of its life-cycle, XWin-NMR did contain a huge hole and nobody used to complain. I wonder if Bruker is a cult more than a instruments-maker. Whatever Bruker does is forgiven by its loyal customers. In the case I am writing about, you had a value of temperature stored in your file that had nothing to do with the actual temperature at which the spectrum was acquired. For example: the probe was at 240 K and the software reported the value 298 K. The reason was obvious, almost natural: the temperature could be controlled by Win-NMR, but in practice you were (and still are) using another program, called "edte", which operates directly on the thermal unit and can't be aware of the active data-set, or of the existence of a spectrometer, for that matter. With edte you can set and monitor the temperature of the probe, but the value is not saved in the spectrum. (The latter belonging to XWin-NMR, not to edte).
It was unbelievable that at the same time Bruker could claim to care about GLPs. The situation went on, in this fashion, for many years. The last generation of XWin-NMR solved the problem: when the acquisition is completed, the software reads the temperature of the probe and stores this value into the spectrum.
This anecdote tells why old Bruker users admire the successor TopSpin. The cause of their happiness is the end of their past sufferings.
Monday, December 11, 2006
I ignore the history of Acorn and generally don't like to express my thoughts about people. Unfortunately I assigned to myself the task of defeating the generalized laziness of the web and now I am forced to describe Acorn's main product, called NUTS (NMR Utility Transform Software). The bad thing is not NUTS, but the laziness of the web. You can find countless lists of NMR software whose only effect is a great loss of time for all their readers. Most of the programs you can find have serious limitations and can be considered unusable. The compilers of the list never find the time to download, test the programs, write a review and discourage you from doing the same.
This particular case (NUTS) is a borderline one. The program is worth to see because is a living fossil, just like its nice vintage icon suggests. Using NUTS can be as funny as visiting a museum. The web site is informative, readable, moderately vintage (by today's web standards) and certainly a valuable source of information. I appreciate a lot the simplicity of the pricing policy, its transparency, the immediate availability of a downloadable demo, etc.. a lot of things that I am in tune with.
There is nothing wrong in writing a program in an ancient DOS-like style. It's the right answer to the hundreds of icons of Topspin (there are no icons into Nuts). I am seriously concerned, however, at the mere idea that in 2006 a chemist could invest money in buying such a fossil. The user can only open a single window with a single pane and two spectra cannot remain open simultaneously. There is something more unusual, though. In a single word: this program is modal, if you know what it means. Before testing Nuts I believed that all Bruker software was modal. Today I must say that Bruker software is 40% modal, while NUTS is pure modal. It's so modal that the manual doesn't even mention the existence of modes and calls them "subroutines". There are 9 of them, plus the starting mode, called "base level". In simpler words: there is a total of 10 different interfaces (just like ten different programs) and the user is expected to learn them all. I wonder how somebody can remember where the commands are because, even when the menus bear familiar names like "Edit" and "View", the menu items are highly non-standard. It's not a matter of opinion, is a matter of design. I mean: if you are writing a program for the Macintosh, like Acorn is doing, you should adhere to clearly stated principles and practices. The principle is that half of your commands should be the familiar "Copy", "Paste", "Undo"... in their familiar menu positions.
The good thing is that you can just ignore the menus. All the NUTS command can and should be written in the command line. Menus have an aesthetic function. In the basic level you find six equivalent menu commands: "Processing Parameters", "Set LB", "Apply EM", "Apply GM", "Set Sine Phase", "Apply Sine Multiply". They all open the same dialog window!
A noteworthy property of most menu items is that they are not context sensitive. They are never dimmed, even when there is no spectrum "on stage". In practice, half of the commands, if given at the start of the program, crash the program itself! It's also funny to see that they make no distinction of FID or spectrum. Knowing "what's on" is 100% responsibility of the user. The sine-bell multiplication, for example, that is intended for FIDs, can easily be applied to transformed spectra, even if I can't find no good reason for such an operation.
All the above things are today considered "non-professional", yet they represent surmountable defects. The average user can discriminate by himself if he is working on a FID or on a spectrum and he will soon learn to do without menus and use the command line exclusively. Now you also know why there is no "Undo" command: the program crashes as soon as you make a mistake, nobody had ever the opportunity of using "Undo" and it was eventually removed. Using the same logic, however, most of the remaining menu items could have removed: For example, the first 6 items of the "Process" menu, plus the six other items already cited, could all have been put into a single dialog.
In my exploration I haven't found the commands to enhance/decrease the apparent intensity of the spectrum. It may be a limitation of the demo version, but at this point I am full of doubts and believe in nothing. Is NUTS actually "usable"? The web site says that they are developing the Mac version because they received specific requests in that direction. How is it possible that somebody who loves the Mac can order a program driven by the command line? How is it possible that they sell NUTS 1D at 499 USD and the fully fledged Mestre-C sells at 265 euro only?
This review refers to wxNuts version 0.6.0, compiled on October 19, 2006.
Topspin NMR free download
Did you know there is a long list of NMR programs that can be freely downloaded from the internet and work with your Bruker files? Check them out!
Other free programs are available upon request (with some restrictions):
Saturday, December 09, 2006
Building your personal NMR database couldn't be simpler. All you have to do is to process your spectra as always. They are automatically indexed by Spotlight and can be searched either with Spotlight or with Speclight (free).
It has always been possible, on spectrometer of any brand, to store a "title" that describes the sample and the experiment. iNMR automatically imports that text and feeds Spotlight with it and with other notes you insert with the acclaimed iNMR annotation tools. To increase your searching options you can also:
- Include the elemental formula of your compounds (and of your impurities too!). If you include this information you can perform searches based on the molecular weight, the number of carbon atoms, etc...
- Include the SMILES strings of your compounds (it's a simple matter of Copying & Pasting). If you include this information you can perform searches based on molecular fragments.
- Perform peak-picking. If you are not used to it, start today. Even if you hide the output, you'll be able to find peaks wherever they are, simply typing a range of chemical shifts. Very handy to recognize impurities!
All your spectra will be indexed, old and new ones, after you install iNMR 1.6. You can also use Speclight to retrieve chemical files (indexed by ChemSpotlight).
Monday, December 04, 2006
Speclight is a free application, with apparently no precedent, that retrieves NMR spectra and other documents in the file system (even on a network). Requires Mac OS X Tiger and, possibly, iNMR 1.6. When I say "other documents" I mean those indexed by ChemSpotlight. Both the cost and the effort of building a personal NMR database are suddenly reduced to zero. Speclight will debut next week; source code available upon request.