On Mon, Apr 11, 2022 at 3:02 PM Michael Quigley
<MichaelQuigley@xxxxxxxxxx> wrote:
Yes, you have to understand the methodology used to determine language "popularity."
This is much easier said than done. I realize a lot of people *think*
they know how the TIOBE index works, but... I wouldn't be so sure.
Case in point:
In part, it's based on frequency of searches.
In part? I think I know why plenty of reasonable people would come to
this conclusion. On the home page (
https://www.tiobe.com/tiobe-index/)
it says:
<blockquote>
The ratings are based on the number of skilled engineers world-wide,
courses and third party vendors. Popular search engines such as
Google, Bing, Yahoo!, Wikipedia, Amazon, YouTube and Baidu are used to
calculate the ratings.
</blockquote>
That verbiage could easily be interpreted as saying that the number of
skilled engineers, courses, and vendors play at least some part in the
calculation.
But, even assuming that they do, how would TIOBE determine those
things? Could it be that those are *implied* by the number of hits on
search engines?
The definition page
(
https://www.tiobe.com/tiobe-index/programming-languages-definition/)
says:
<blockquote>
Basically the calculation comes down to counting hits for the search query
+"<language> programming"
</blockquote>
And then they try to give some details behind that. That page provides
all the documentation on the TIOBE index methodology that most people
have access to.
I think it's pretty clear the *entire* calculation, if that definition
page is to be believed, consists of counting all the qualifying search
engine hits for all the qualifying programming languages, and then
figuring out the share of that total for each language.
And even that seemingly simple and straightforward calculation is
fraught with things that make one wonder. For example, the home page
says that Bing is one of the search engines used. Sounds reasonable.
But then the definition page says (as of this writing; surely subject
to change) that Bing.com does NOT qualify! Why? Well, they give
clearly computerese "reason codes" for all the search engines that
didn't qualify, and Bing's reason code is "NO_RESULTS_AT_ALL". What?
Seriously?
So, Bing.com doesn't qualify, but Ebay.com and Etsy.com do. Riiiiiight.
(I get that Ebay's inclusion is predicated on the fact that lots of
folks are selling computer programming books. And if you try, for
example, +"go programming" on Ebay, you do get mostly books on the Go
programming language. But also on my first page of Ebay hits, I got
"Signia Pure Charge & GO T 1AX + Telecoil RIC - Rechargeable
+Programming+Charger" for $575. (Those are hearing aids, for those who
don't know.) And this: "Murach's Python Programming Michael Urban,
Joel Murach Paperback Used - Very Go" (which is a book on Python in
"very good condition" with some of the description cut off). For a
laugh, try that Go query in Etsy's search box. Then, to actually get
programming-related merchandise, try +"golang programming" on Etsy,
but be prepared for lots and lots of mugs/shirts/etc. featuring only
Python, or only JavaScript, or a bunch of programming languages none
of which are Go, or programming humor that doesn't mention any
specific language at all.)
Even assuming "popularity" is something that is rigorously definable,
and even assuming it is something objectively measurable, and even
assuming it is something worth measuring, TIOBE just fails really,
really hard on all counts.
With the giant proviso that all three of those assumptions I just
listed are HIGHLY QUESTIONABLE, I would say there is clearly more
value in other sources, such as the PYPL index, or data gleaned from
Stack Overflow or GitHub.
John Y.
midrange.com
