OriginLab has created a standard for rich transmission of technical graphics, argues Felix Grant
Three views of a sample Origin project – in Origin 7.5 itself (background left), in the Origin Viewer with data partially highlighted for extraction (top right foreground) and, again in the viewer, showing a graphic representation (bottom right foreground).
Standards evolve slowly, painfully and frustratingly. There's never any certainty, until after the event, which they will be; they are not always the best engineered (VHS vs Betamax, anyone?); and they are often only partially adopted. Nevertheless, trying to spot them and discuss them in advance is worth doing – if only because a good candidate has a better chance of adoption if a lot of people are talking about it. OriginLab Corporation recently released one such candidate which, in my opinion, deserves to be talked about: the freely distributable Origin Viewer.
OriginLab's publicity people describe this as being similar to Adobe Acrobat's Portable Document Format (PDF), and up to a point they are right. It could also be reasonably compared to the many viewers and plug-ins that have been released by other software publishers, since it makes content in Origin's own project (OPJ) file format available to those who do not have Origin itself. So far so useful, but neither comparison really does justice to the possibilities here.
One of the things that the desktop scientific computing world lacks, at present, is a standard for rich transmission of technical graphics. It's easy enough to send a graphic, of course; just save it out to JPG, TIF, or whatever. The raw data from which it was generated can be sent as an accompanying text file, spreadsheet or database – though often isn't, or gets lost along the way in repeated onward forwardings from recipient to recipient. Documentation can be in the data file, but it is usually a separate attachment. The whole lot could, of course, be bundled up as a word processor file (most recipients can read Word, even if they don't have the program itself), or a PDF – but the result is not easily reusable. A spreadsheet (Excel has similar ubiquity to Word) is the only option at the moment, and it's a clunky one. One of the delights of Design Science's MathType is the fact that all the data for recreating, editing and reusing content is encapsulated within (and retrievable from) the universally readable GIF graphics file used to display it. Wouldn't it be nice if a technical graphic could encapsulate its data and documentation in a similar way?
That, essentially, is what Origin Viewer does. Sitting comfortably in the corner of a USB flash drive, if required, it takes an OPJ file and displays all of its content: multiple datasets, multiple compound graphics, and documentation. The data can be extracted (with full numerical precision) for analysis in any appropriate worksheet-oriented program you have to hand; I've successfully tried DesignExpert, Excel, Genstat and TableCurve, for example. The graphics can be taken (as a full quality metafile) anywhere that a graphic might be needed: Word, Photoshop, Excel, whatever. Text too, of course. A PDF is a useful thing, as is an Excel workbook, but for passing around scientific data it doesn't begin to approach this for elegance and convenience. And, crucially, there is no reason why that content has to originate in Origin itself.
OriginLab make available for custom application development a software component, Orglab, which can be used to create OPJ files independently of Origin. That means, for instance, that you can export an Excel workbook to an OPJ file and pass it around with the viewer. A sample Excel macro is provided with the viewer download for demonstration purposes, but it's not dressed to impress and I think OriginLab is missing a trick here.
When I asked whether other applications would be allowed to add OPJ to their export or File Save As menus using Orglab, I was told 'we would love that' and it's easy to see why. Widespread adoption of the OPJ as a distribution format would give Origin a natural edge as the obvious choice for extended graphic handling of the content. There was doubt that any software publisher would follow this route since Origin would be seen as a competitor, but a survey of other products doesn't seem to suggest that this is so. Most analysis and visualisation packages offer a range of export formats to some degree, and a common packaged export format could offer them a lot of advantages. I am not at all sure that Origin is really a competitor to (for example) Excel. On the contrary, it is a complementary tool – as is recognised by the strong Excel communication at a worksheet level.
The natural follow-on to release of the viewer, it seems to me, would be provision of a really slick, ready-made file save module – first for Excel, then perhaps for other strategic products – also for free distribution. OriginLab sees this as something for the user to take on, but I beg to differ on several grounds. Most users don't have the depth of application-specific programming experience to do the job as well as OriginLab could do it, for one thing. A plethora of different interfaces to the same task in the same application isn't conducive to either standardisation or uptake for another.
Even better than such specific modules would be a 'distiller', analogous to Adobe's Acrobat software or the various third party equivalents. As a heavy user and distributor of graphics-fronted reports, I would pay good money for such a distiller. At present, I routinely output PDF instead of paper or application-specific files for most purposes. I estimate that OPJs for this viewer would replace about 60 to 70 per cent of my present PDF output from data analysis if there were an easy way to produce them.
It only supports the Wintel platform so far, and there are data set size limits, but this viewer is a really good idea that has gotten me very excited. If it catches on, it could become a valuable lingua franca. Good, solid, consistent links to other strategic products, either through specific export modules or a generic standalone converter, would greatly increase its chances of deservedly becoming an accepted distribution standard – and, I believe, greatly benefit OriginLab at the same time.