I’ve never seen such a content-less post on that website. “Hey, they have no Perl 6 data so I’m just gonna guess how many % of the Perl 5+6 are Perl 6 and then from the next sentence on I’m gonna ignore Perl and plug some languages that are low on the TIOBE index but they’te totally cool!111” WTF
Certainly not content-less. Sure, it’s a bit meandering stream-of-consciousness before the title gets addressed, but the idea that maybe ‘cumulative popularity over time’ is more important than ‘current popularity’ seems useful and easy to overlook.
I enjoyed reading it and I don’t see why it is necessary to trash it this way. I’d gladly read a John D. Cook post like this one every day. Most days we even get something better.
See, I usually agree with his writing being above average and maybe I did miss the point you mentioned, but even rereading I don’t really see your ‘cumulative popularity over time’ is more important than ‘current popularity’’ really addressed.
Sorry for the harsh language, I’m extra critical of any extrapolation of data based on a) TIOBE b) GitHub stars, and c) questions asked on StackOverflow - unless there’s a huge caveat or you’re simply looking for popularity - but deducting any kind of viability from those metrics (compared to others) is a moot point.
In any way, thanks for comment that did provide more insight than reading it twice, but from the amount of upvotes I can only deduct that at least 9 people agreed :)
Well, to put it in context, when you link through to the TIOBE Index, it reads April Headline: Perl is having a hard time, so it’s natural to address Perl.
The only thing I dislike about the article is that he vaguely states
Common Lisp, Erlang, and F# would all be safer bets for a production software project than several more popular languages.
but doesn’t specifically name them. So what languages would those be? Well the TIOBE itself claims Objective-C is having a similarly hard time since it’s basically being replaced with Swift. What else: C++? VB.NET? PHP? Delphi? VB6? Assembly?? At some point, the list is just not homogeneous enough to make meaningful viability statements. Who’s going to write back-end code in MATLAB? (even if it’s popular)
Overall, I agree with the author: most languages will do just fine for most sorts of work. I mean, large parts of lots of things are written in bash. It’s not necessary to always reach for the “IBM” of languages, the non-risk never-gonna-get-fired choice. (“Blub—it’s what everybody else is using”)
I work with a large amount of REXX and Object REXX and other “unpopular” language code which I have no qualms about. Perhaps an actual non-subjective measurement of risk might be determined or at least conjectured and stated?
I’d imagine there is comparatively very little risk in unpopular but old and established languages while there is a high risk of new and unpopular languages becoming “unviable”.
The problem is that this article gives very little thought to formally defining its criteria in an objective and measurable way.
I’ve never seen such a content-less post on that website. “Hey, they have no Perl 6 data so I’m just gonna guess how many % of the Perl 5+6 are Perl 6 and then from the next sentence on I’m gonna ignore Perl and plug some languages that are low on the TIOBE index but they’te totally cool!111” WTF
Certainly not content-less. Sure, it’s a bit meandering stream-of-consciousness before the title gets addressed, but the idea that maybe ‘cumulative popularity over time’ is more important than ‘current popularity’ seems useful and easy to overlook.
I enjoyed reading it and I don’t see why it is necessary to trash it this way. I’d gladly read a John D. Cook post like this one every day. Most days we even get something better.
See, I usually agree with his writing being above average and maybe I did miss the point you mentioned, but even rereading I don’t really see your ‘cumulative popularity over time’ is more important than ‘current popularity’’ really addressed. Sorry for the harsh language, I’m extra critical of any extrapolation of data based on a) TIOBE b) GitHub stars, and c) questions asked on StackOverflow - unless there’s a huge caveat or you’re simply looking for popularity - but deducting any kind of viability from those metrics (compared to others) is a moot point.
In any way, thanks for comment that did provide more insight than reading it twice, but from the amount of upvotes I can only deduct that at least 9 people agreed :)
Well, to put it in context, when you link through to the TIOBE Index, it reads April Headline: Perl is having a hard time, so it’s natural to address Perl.
The only thing I dislike about the article is that he vaguely states
but doesn’t specifically name them. So what languages would those be? Well the TIOBE itself claims Objective-C is having a similarly hard time since it’s basically being replaced with Swift. What else: C++? VB.NET? PHP? Delphi? VB6? Assembly?? At some point, the list is just not homogeneous enough to make meaningful viability statements. Who’s going to write back-end code in MATLAB? (even if it’s popular)
Overall, I agree with the author: most languages will do just fine for most sorts of work. I mean, large parts of lots of things are written in bash. It’s not necessary to always reach for the “IBM” of languages, the non-risk never-gonna-get-fired choice. (“Blub—it’s what everybody else is using”)
I work with a large amount of REXX and Object REXX and other “unpopular” language code which I have no qualms about. Perhaps an actual non-subjective measurement of risk might be determined or at least conjectured and stated?
I’d imagine there is comparatively very little risk in unpopular but old and established languages while there is a high risk of new and unpopular languages becoming “unviable”.
The problem is that this article gives very little thought to formally defining its criteria in an objective and measurable way.