Shared publicly  - 
 
I want to thank everyone who contributed to Go 1. We have over 200 external contributors, some of them quite prolific, and Go is better for it. Thank you.

From the very beginning, we knew that Go had to be an open source project, free for anyone to use, adapt, and improve upon. That's why we're so excited to have such great contributors.

Even though Go is designed to address the software development problems we have at Google - and is now used in App Engine, YouTube, and other places within Google - we understand that they are faced by many people and that for Go to be a true success, it must have the broader applicability and appeal required to stand on its own. That's why we're so excited to see companies like Atlassian, Canonical, Heroku, and Tinkercad using Go too.

We've understood from the initial launch that Go needed to be more than just Google's language. The front page of golang.org has never mentioned Google and carries none of the Google logo or branding, and that is not accidental.

In some of the coverage of Go 1, I see it repeated that Go is dropping in popularity because TIOBE says so. This is not what I see. If anything the Go world seems busier than ever, with more mailing list traffic than I can read and some half dozen books out or out soon.

If you read TIOBE's methodology carefully [1], you can see why they think Go is dropping in popularity. They judge "popularity" by averaging result counts from different search engines for "<name> programming": "C programming", "Java programming", and so on. However, they have judged that "Go programming" has too many false positives, so for Go they use "Go programming" "Google", requiring both terms. In this way, TIOBE essentially ignores all the references to Go as an open source project. Requiring "Google" to appear with "Go programming" for a hit to count is like requiring "Bell Labs" to appear with "C programming" or "Sun" to appear with "Java programming". The pages that talk about Go programming and feel it necessary to mention Google are mainly news articles, not actual discussions of the language.

What we learn from TIOBE then, is that Google Go is fading in popularity. That's great: it means the open source Go - the real one - is winning.

[1] http://www.tiobe.com/index.php/content/paperinfo/tpci/tpci_definition.htm
97
22
Matt Horsnell's profile photoRuss Cox's profile photoJemes Hsu's profile photoRoberto Bayardo's profile photo
18 comments
 
Regarding TIOBE... it shows the importance of choosing a good name if you want it to be easily found by search by the users. Try to search Learn Go and see struggle in the results that have nothing to do with the programming language https://www.google.com/search?q=learn%20go
Russ Cox
+
1
3
4
3
 
[learn go language] and [learn go programming] both work fine. Even [learn go] shows something about Go the language on the 4th hit on Google. This is not really any different than naming your company 'apple' or 'sun' or your product 'windows'. People can disambiguate, and search engines are getting better at it too. Search engine behavior need not dictate naming.
 
Sometimes that works but not always. The domain name is golang; the language is Go. Certainly "golang programming" is not a common phrase.
 
Agree, but at least "golang" has more results (on google) than "go programming" +"google" :)
 
I wish I had been able to contribute more to go, but thanks to the go team for having such a great contribution and review process and a code base that's a delight to work with.
 
I always search by golang and have gotten good results.
 
+Russ Cox 'apple', 'sun' or 'windows' were named like that way before to internet era, and now they are to big to have search problems. Search engine behavior might not dictate which name you choose but if you have complete freedom (like in this case) there are some problems that are just too easy to avoid. I was somehow still expecting a rebranding to golang for the 1.0 release, seems not.
 
I have never had any issues searching for Go related material -- I mostly use the keyword "golang". However, I would argue that measuring popularity by search results of the name is a very poor and misleading measure. Maybe if it looked at things like the number of github projects with that language used and compared it historically, you could actual get an interesting sample of growth of popularity.
 
congrats, your team is truly an inspiration!
 
I always imagined most search hits were posts of people having problems. "My C program crashes, why?" "I'm using all these things in my Java program, how do I make them fit together?" Maybe people just don't have problems with Go.
 
+Russ Cox My problems were in queries like "Go mysql", or "Go algebraic data type", it often doesn't give useful result. Searching golang sometimes helps, since some people are kind enough to mention this keyword in their projects, and since it implies searching the golang-nuts mailing list. (IIRC google made an exception for C# in order to get better results for the language).
 
I guess programmer mindshare is the utmost importance factor for Golang to be adopted mainstream. One area that strikes me is the Github ranking (https://github.com/languages/Go) which is used by some programing popularity index site. When I looked at Golang about 6 months - 1 year ago, it was #29, and it is now #31. My guess is this has a lot to do with more projects are in Google Codes now rather than Github. But some programmers will take the Github ranking and sites such as TIOBE to mean something.
As for making Go more search friendly, why not just make Golang the official alias for "Go Programming Language" and use Golang everywhere in official doc and site? One area to push for more Golang adoption is to make Golang first class language for Android native app development. I hope for such announcement in Google I/O this year. For many, if Golang is not mainstream, it is hard to convince coworkers/managers to adopt.
 
Who cares for popularity rankings anyway? Good things are hardly ever adopted immediately/fast. Go is by far the most pleasuring language I've ever worked with and that fact is not influenced by popularity (especially because the std library is that good).
 
Now that Go1 is here, I expect go numbers to go up by an order of magnitude. I believe that once Go1 was planned for early 2012, a lot of people like myself put the brakes on using go only to let the language stabilize. It was hard for many to keep up from last autumn until yesterday, especially if one used third party packages.

Now that I know that the code I write today will compile for years without worrying about some packages breaking or about a change in the standard library API, I for one will commit to use go for other purposes than personal projects. And I expect many others to do so. I think we'll see good numbers for go in github, stackoverflow and such, excluding TIOBE because of aforementioned issues.
 
My only go related wish is that the std package documentation could be more easily navigated. Is there an emacs godoc plugin?
 
Awesome. Now hurry up and work on Go 2, so someone can write a paper named "Go 2 Considered Harmful" (and a rebuttal, "Structured Programming with Go 2").
Add a comment...