Why is Go doing so poorly in the Tiobe index?

[Peter Kleiweg]

Go isn't even in the top 50 anymore, and there are some really
obscure languages in the top 50. Is Go really such a tiny
island? Only three years ago Go was language of the year. Where
is it all gone?


http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html


--
Peter Kleiweg
my Go cookbook: http://www.let.rug.nl/~kleiweg/go/

[DisposaBoy]

On Sunday, June 10, 2012 11:50:50 AM UTC+1, Peter Kleiweg wrote:

Go isn't even in the top 50 anymore, and there are some really
obscure languages in the top 50. Is Go really such a tiny
island? Only three years ago Go was language of the year. Where
is it all gone?


http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html

I don't know the *why* but I can offer some assumptions.

1. Go isn't the latest things, IIRC tiobe is based on what the search engines are saying... i.e. the latest new thing gets closer to the top

2. Go isn't like many other languages where you can't go without a search engine on the ready. This is due to the high quality of the docs, the fact that there are dedicated services out there for documentation on packages both std and third part, so people are more like to just use those as opposed to a search engine and let's not forget godoc

I'm sure there are probably many more reasons but for me those are the obvious ones

 

[David Symonds]

Your subject line has a faulty premise. You should be asking why the
TIOBE index is not indicative of reality, as it's largely irrelevant
how Go is performing in it, just as it's fairly irrelevant how well Go
is performing in the shootout.alioth.debian.org benchmark game.


Dave.

[Christoph Hack]

It's probably not that easy to judge the success of a programming language

by the number of search results of a specific query. According to the definition

of the Tiobe Index, they are currently ranking Go by the maximum number

of search results of either of those queries:

- "Go Programming"

- "Google Go"

- golang

The term "Google Go" is rarely used in general and the term "golang" (which

is my personal favorite and probably the only one that I am using regularly) is

mostly used on social networking sites which are often not crawled by search

engines. So, the first term is probably the only one that yields any significant

amount of results.

But if you are unhappy with the result, just look at the Transparent Programming

Language Index at http://lang-index.sourceforge.net/. If I'm not mistaken, they

currently just search for "Go", which results in the 16th place (and the 10th place if

you just take general-purpose programming languages into account). A much

better result. :-)

-christoph

[bugpowder]

So, are all external metrics irrelevant?

This is a strange way to think about "reality". 

[David Symonds]

On Sun, Jun 10, 2012 at 10:56 PM, bugpowder <mit...@gmail.com> wrote:

> On Sunday, June 10, 2012 7:31:19 PM UTC+8, David Symonds wrote:
>
>> Your subject line has a faulty premise. You should be asking why the
>> TIOBE index is not indicative of reality, as it's largely irrelevant
>> how Go is performing in it, just as it's fairly irrelevant how well Go
>> is performing in the shootout.alioth.debian.org benchmark game.
>
>
> So, are all external metrics irrelevant?

No, just inaccurate metrics. There's other ways to measure programming
language adoption that are more accurate, and are probably relevant to
gauging Go's success or otherwise.


Dave.

[Rodrigo Moraes]

On Jun 10, 8:42 am, Christoph Hack wrote:
> But if you are unhappy with the result, just look at the Transparent
> Programming Language Index at http://lang-index.sourceforge.net/.

I didn't know that one. These things shouldn't be taken too seriously,
but they are kinda fun to watch.

-- rodrigo

[Michael Jones]

No at all. The dismissive comment was about metrics more than external. 

It is simply the truth that each scientific experiment has its own power to inform. Now for programming languages, a good experiment would be one that gives insight into how many developers are using the language, how quickly that number is changing (and in which direction!), the nature of the projects undertaken (play vs work vs mission-critical, team size), matters of capability, richness, utility, performance, clarity, "debugability", and maintainability, and the developer's retrospective measure of success, and from this, a sense of what is right and wrong with the language. This would be great to know and is inherently an external measure.

The issue is that looking at search engine queries, especially for a term like 'go' (as opposed to 'FORTRAN' or 'Scala') is not likely to give you a strong sense of the above. Likewise, it cannot be the goal of any serious development team simply to be known, in the Lady GaGa sense, as opposed to being better, in the work hard on all of the above. As such, all language developers are likely to be pleased when such surveys notice them and unconcerned when they don't.

On Sun, Jun 10, 2012 at 5:56 AM, bugpowder <mit...@gmail.com> wrote:

So, are all external metrics irrelevant?

This is a strange way to think about "reality". 

Michael T. Jones | Chief Technology Advocate  | m...@google.com |  +1 650-335-5765

[André Moraes]

> Go isn't even in the top 50 anymore, and there are some really
> obscure languages in the top 50. Is Go really such a tiny
> island? Only three years ago Go was language of the year. Where
> is it all gone?

I think this is because Tiobe index don't count access to golang.org,
golang-nuts, golang-dev.

This list just keep growing:
http://go-lang.cat-v.org/organizations-using-go

--
André Moraes
http://amoraes.info

[andrey mirtchovski]

here's a somewhat official explanation:

https://plus.google.com/116810148281701144465/posts/ExP5BeXUujn

[Andrew]

i usually use somewhat different approach to get a feel on popularity
of the tool/technology - monster.com -> golang
the results are (unfortunately) pretty consistent - neither today nor
1.5 years ago . But the linkedin has one job for go developer.
I think that the push into development of the client side apps might
add significantly to the popularity of the language. At the same time
the R1 was released basically yesterday, so the expectations of
enormous popularity should be given a reality check.

My personal enthusiasm is based on 2 factors:
a. simple, logical syntax
b. ability to use the power of parallel processing with short
learning curve.

all-in-all i hope to see increase in AZ index :) soon.

[Rich]

The tiobe index means nothing to me, I don't program in go because it's popular. I am a sys admin and I don't usually have a ton of time to write programs or wait for them to execute. Go is an awesome tool for me because it has a clean and consistent syntax that makes it easy to write that quick program to sift through that five thousand line of XML file, or synchronize directories where rync wouldn't do the job. I recently wrote a nice short program in an afternoon to verify SSL keys for servers that sit behind a load balancer. I could do that in perl, ruby, or python -- given I had the right extension, stock Go came with everything needed.

[John Barham]

AFAIK the standard Python documentation package on Linux does not
include a search engine. godoc does. Ergo when I need to look up the
docs when I'm programming in Python the first think I'll do is check
an external engine, which TIOBE sees, but for Go I'll first check
localhost:6060, where godoc is running, which TIOBE doesn't see.

John

[GreatOdinsRaven]

First, Go is a young language, really young. 

How long did it take for Ruby to gain any form of popularity? Most of the programming community hadn't even heard of it until Rails came along. Rails was and still is, the "killer app" for Ruby. I remember reading an old Rails book and it had to make a specific point in its introduction that Ruby existed outside of Rails as its own language. 

Second, Go is not just a "web language" - this isn't just PHP in a different skin. It's also a low-level language, yet the audience that could benefit the most from it also one that is extremely hostile to it - the C/C++ crowd. Those guys are in general very traditional and conservative and approach any idiom/idea/way of thinking that is not theirs with extreme hostility and suspicion.

Too low-level for PHP users, too different for the C/C++ crowd, not functional/esoteric enough for the Haskell crowd. Go does so many things so very well yet leaves out at least 1 "favorite feature" from each of the existing camps to scare them away. Until they try it, of course :-)

For web apps, difficult to find hosting. For native apps - other languages are too entrenched with lots of libraries/tools/momentum behind them. 

So what is Go's niche? Where is it used? What is its "killer app"? I'm sure all of these will be resolved with time and it should emerge as a popular language. There really is no reason not to! Any current short-comings (cgo, Windows support, 32-bit GC) are implementation weaknesses, not language weaknesses, which means they can be improved upon in a non-disruptive manner. The future for Golang is bright, I'm betting on it. 

[Paul Borman]

I view Go's "killer app" being fast development and reduced maintenance costs.  If people don't value those then, well, I really don't know what to think of them :-)

Sadly, like plumbing and infrastructure, it is often ignored until it hurts.

From my perspective the single largest inhibitor to using Go is interaction with existing infrastructure.  This is a solvable problem.  For C/C++ libraries you can SWIG.  For all libraries you can re-implement.  People do balk at the idea of re-implementing, however, I believe that to be a short term pain.  When an existing Python or C++ project becomes extremely difficult to maintain it is ripe for a re-write in Go as a re-write is often the only solution to that problem.

    -Paul 

PS:  As for Tiobe, I view it as irrelevant.  There are statistics and there is the truth.  They are not always related.

[Fabio Kaminski]

As for the popularity of a programming language.. i think the
darwinistic way we are used to think about things.. in terms of
"best, better and fitness" .. are not really true for the popularity
of a language..

from my subjective observation.. the most popular and used languages
.. are not always the best fit.. or the ease ones..
but the one bundled with a PLATFORM

Programmers, companies, and almost everybody else.. target a
platform.. and thats why languages as
C# or Objective-C, and (ouch) Javascript are so popular.. because
people want/need theyr programs running on that platform
Windows, IOS and Web..

As for the Go "fight arena".. linux/unix ..and servers.. it may be a
very popular language already..
for me.. im not that lucky.. i love Go.. but i still need to use C++

A good metric about how Go is going, would be some metric targeting
server and unix programming specifically.. as the other platforms are
huge, comparing with the server/infrastructure side ones..

But the we cant match the quality of a server/infrastructure platform
enginner to a web platform developer .. can we? :)

So i think.. sometimes is not popularity that matters.. but the
quality, and the result of the work you gotta do..using that tool..
in the end of the day.. we need to be efficient, and make the things
work the best way possible.. and Go is a very strong candidate
for that matter