0

In my field of research, I have observed that many people are reinventing the wheel for certain types of applications which does not seem to be a great use of research time. To make this more concrete, I have noticed that there are several different file formats to exchange CAD (Computer Aided Design) files which are all not necessarily useful in my area of research (Computational Engineering, i.e. simulation of some sort based on the CAD file). For example, in my area, I mainly need the surface definition of the CAD file which will then be used as a boundary condition in my simulations, yet, any available CAD format makes this process either unnecessarily complicated or straight-up impossible, due to surface impurities that arise in the process of generating the surface information (tessellation, in this case).

That being said (and talking to commercial software vendors who face the same issue), I figured it would be good to come up with a new file format that is specific for our computational modelling requirements and stores surface information in a way that can be directly consumed by our simulation tools. There is an argument to be had about whether we really need another file format (which I don't want to dwell on here), but if I wanted to work on something like this, one thing I noted is that there is no standarisation in place.

As an example, a popular file format we commonly use is the *.stl file format which is used for 3D printing. Its great for that 3D printing, it stores the surface definition and even if there are impurities, these will be so small that the 3D printer will basically just print over them. In my area, even a tiny gap in the surface will cause issues setting up simulations, so time-consuming pre-processing has to be done to bring the surface quality up to a standard required by the simulation tools. Even more problematic is that the company that came up with this format does not seem to be around anymore (or at least it is no longer associating itself with the file format) so basically the file format is not standarised, everyone can have their own implementation of it which leads to some incompatibilities when trying to generate it with one software and then reading it with another.

Thus, given the lack of a standard, it would be useful to have a standard for a domain-specific file format. This would be mainly used (initially) for academic research, though a new file format (and its standard) may not qualify as novelity and thus I wonder how one would go about publishing such a standard in a journal? Has anyone experience with such an endeavor?

To be clear, initially I am not considering going for an ISO standarisation or similar (due to cost implications) but rather follow best practices such as guidelines provided by the open-source initiative: https://opensource.org/osr

The other reason I am looking into publishing this is to make this available to academic researchers who would likely be the early adopters and it makes sense to disseminate that knowledge in areas which the target audience is reading.

Looking at similar efforts elsewhere, there seems to be a trend to publish efforts in conferences (which may be a good idea in terms of informing other researchers about this development), but when it comes to a standard itself, it feels more natural to put this into a journal rather than a conference (i.e. this would be a reference paper).

Any suggestions on how to go about publishing a standard would be welcome. In my field, as far as I know, there are no such journals dealing with standards, perhaps there are journals dealing with this specifically?

7
  • 5
    Unfortunately, this is what you face: xkcd.com/927
    – Buffy
    Aug 24, 2022 at 11:24
  • there is some truth to it but as I was mentioning above, the most used file format does not have a standard but rather a set of beliefs how this should be implemented (hence the idea of standarising a bespoke file format).
    – tom
    Aug 24, 2022 at 12:43
  • 1
    Knowing what my engineers do while converting 3D CAD files into input for other analyses, it is highly unlikely that your file format would cover any single use case, much less all the use cases.
    – Jon Custer
    Aug 24, 2022 at 13:53
  • Not my field, but wouldn't it be more useful to provide (automated) methods to process/correct commonly used file formats for analysis rather than contributing yet another file format? Admittedly this is probably also harder to do.
    – atom44
    Aug 24, 2022 at 14:28
  • A standard does not usually become a standard through the work of one person. (Representatives of) the community need to agree that it should be the standard. Publishing a proposal for a standard can be part of the process but it works best if it is the last step and not the first of reaching sufficient buy-in.
    – Roland
    Aug 25, 2022 at 4:46

1 Answer 1

2

Your approach and objectives are not very compatible with the traditional model of publishing in journals.

Said model is that a researcher makes an advance in a specialized area, writes up a compelling article and sends it to experts who decide whether the advance merits being shared.

Contrast this with what you've described. What you are proposing is not an advance but a restriction in how a certain type of file should be prepared. People use it in different ways; you believe that there should be fewer ways to use it. This already demands a very strong reason for further consideration. Your description does not make a compelling case- you don't indicate that other researchers face problems with this. Commercial software vendors may have some agreement with your idea, but orderly commercial productions are often orthogonal to the messy and diverse requirements of researchers. That's why many researchers write programs, despite not being trained or compensated for it.

Further, the standard would work if and only if you understand the needs of most researchers using these types of files. You must judge whether you are in a position to assert this; the information in the question does not suggest this. Having worked with these types of files for different purposes, I've always found tools (mostly written by researchers) to help me with converting and making different versions compatible. I cannot claim that all purposes are met by such methods though.

Finally, if a researcher must do this, conferences are the right way to go. There one can get a real sense of whether the idea is useful and worth exploring, as well as finding stakeholders who would support such an effort in the future.

3
  • Well, my question was about how to publish a standard in an academic context, not whether I have the credentials to undertake such work. I was sparse on the information I provided because it was not adding value to the question. That said, if you wanted to have more info, there are essentially 3 criteria a file format should have to be useful in our field (CFD): support for higher-order, tessellated elements (quadratic/cubic/quartic triangles), water tightness guarantees and only manifold connected triangles.
    – tom
    Aug 26, 2022 at 8:23
  • Sure, tools can help to some extend, but you can't transform a linear triangle to a higher order element without losing information. There are similar efforts under-way using a dedicated file format that stores computational meshes (CGNS files), but they don't look at CAD. The idea here is to complement this by having dedicated meshing and CAD file formats which remove a lot of time wasted on pre-processing (honestly that research time and money can be spent in better ways). That being said, I probably agree that conferences may be a better starting point to gauge interest
    – tom
    Aug 26, 2022 at 8:26
  • It isn't much about having credentials as being aware of the needs of all users. An alternate framing could be convenience vs customizability-which is a recurring debate in many R&D areas. Besides conferences, maybe CFD support fora/user groups could be a good place to gauge interest. Aug 26, 2022 at 13:13

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .