On 15/01/18 23:06, Anthony Gatlin wrote:
> I currently develop for my "day job" primarily in C# and JavaScript, but
> I really love Prolog and am very interested in using it in a
> professional setting. From everything I read, it seems like most of the
> A.I. and machine learning work these days centers around Python and R. I
> don't see a significant amount of activity in the Prolog community. (I
> am not saying it isn't there. I may just be looking in the wrong place.
> That is part of the motivation for this post.)
Work is this direction is mostly happening around YAP Prolog and mostly
in the life sciences arena AFAIK. Fabrizio has collected a lot of this
stuff for SWI-Prolog on
http://cplint.ml.unife.it/
IMHO we should not start to compete with R and NumPy. Prolog is not
suitable for this, but to some extend this also holds for R and NumPy,
both doing all the real work mostly in C. Technically this would also
suit Prolog. I see little hope we can successfully compete and there
isn't that much point in doing so. It is way easier to provide good
interfaces to these tools. We have that for R (three of them). Prolog is
particularly good at integrating data from multiple sources using
multiple structures (RDBMS, RDF, XML, ...).
We have a case study on data analysis running on SWISH, using Prolog to
abstract away from the raw data, R for statistical analysis, clustering
and producing charts and Prolog again for dealing with graphs. Would be
great to show, but unfortunately the underlying data cannot be
published. Search for "SWISH datalab" and you'll find some of the
publications.
> I would simply like to ask the following questions.
>
> 1) Is Prolog still being used for "new development" in industry or
> academia, or is it simply a wonderful language whose star is fading?
That is a really hard question. I can only talk for SWI-Prolog; I
notably have little clue what happens around SICStus and other
commercial systems. SWI-Prolog is definitely used for new projects,
mostly initiated by startups. Unfortunately, with few exceptions, this
work is usually secret. Next, there are quite a few one/couple of people
businesses using it. There are also some big compagnies using it, though
I think mostly in R&D. This is probably comparable to academic usage.
I've seen a number of startups starting with Prolog and moving to
something else as the business grew due to demands from investors and
lack of qualified Prolog programmers. That is a pity and probably
technically not necessary. Some also claim is was for Prolog that they
managed to open up their market in time with modest resources.
> 2) Are there actually decent jobs/contract projects available that
> include Prolog as either a primary or secondary development tool?
> 3) If one wanted to find an opportunity involving Prolog, what might be
> the best places to look?
These seem related. There seems to be a problem for Prolog programmers
finding jobs as well as for projects to find Prolog programmers :(
> 4) What other skills besides pure Prolog knowledge would companies or
> organizations that hire Prolog developers generally find most attractive?
I guess that is very diverse. I see projects in finance, natural
language, testing, life sciences, robotics, configuration and I'm sure
I'm missing domains. Some are almost entirely Prolog. Most combine
Prolog with other languages where the mechanisms used for combining and
the languages used are diverse. What could be the shared non-Prolog
knowledge required? I can't think of much besides the general skills any
programmer needs.
Cheers --- Jan
> Any guidance or insights are much appreciated.
>
> Thank you in advance for sharing your thoughts.
>
> Anthony
>
> --
> You received this message because you are subscribed to the Google
> Groups "SWI-Prolog" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to
swi-prolog+...@googlegroups.com
> <mailto:
swi-prolog+...@googlegroups.com>.