Happy birthday, come on!
Sunday, we celebrated the 15th anniversary of the release of Go Open Source!
So much has changed since Go's 10 years old anniversaryboth in Go and in the world. In other respects, he therefore remained the same: Go remains attached to stability, security and software engineering and large -scale production.
And it's going strong! The GO user base has more than tripled in the past five years, making it one of the fastest growth languages. Since its beginnings only fifteen years ago, Go has become one of the first 10 language and the language of the modern cloud.
With the versions of Go 1,22 in February And Go 1,23 in AugustIt was the year of for loops. Go 1,22 manufactured variables introduced by for loops iteration parkRather than by loop, addressing a long -standing language “Gotcha”. More than ten years ago, leading to the release of Go 1, the Go team made decisions concerning several language details; Among them if for The loops must create a new loop variable on each iteration. In a funny way, the discussion was quite brief and clearly unconnivided.
Rob Pike closed it like Rob Pike with one word: “Stet” (let him be). And it was like that. Although apparently insignificant at the time, years of production experience underlined the implications of this decision. But during this time, we have also built robust tools to understand the effects of changes to go – in particular the analysis on the scale of the ecosystem and the tests on the entire Google code base and established to work with the community and obtain comments. After tests, an in -depth community analysis and discussion, we have deployed the change, accompanied by a hash bishage tool to help developers identify the code affected by large -scale change.
Change to for Loops was part of a five -year trajectory of measured changes. It would not have been possible without Under -term language compatibility Introduced in Go 1.21. This, in turn, was built on the foundations placed by GO modules, which were introduced in Go 1.14 four and a half years ago.
Go 1.23 still built on this change to introduce iterators and Land -based loops defined by the user. Combined with generics – introduced into Go 1.18, only two and a half years ago! – This creates a powerful and ergonomic base for personalized collections and many other programming models.
These versions have also brought numerous improvements in production for production, especially Make up for the HTTP router of the standard libraryA Total revision of execution tracesAnd stronger random For all GO applications. In addition, the introduction of our First V2 Standard Library Package Establishes a model for the evolution and modernization of the future library.
In the past year, we have also deployed with caution Opt-in telemetry For Go tools. This system will give developers GO data to make better decisions, while remaining completely open and anonymous. Go Télémetrie appeared for the first time in Gopls, the Go language server, where it has already led to a litany of improvements. This effort opens the way to program in an even better experience for everyone.
In the meantime, we evolve to better take advantage of the capacities of current and future equipment.
The equipment has changed a lot in the past 15 years. In order to ensure that Go continues to support the large -scale high -scale production workloads for the following 15 years old, we must adapt to large multicores, to sets of advanced instructions and to the growing importance of the locality in hierarchies of more and more uniform memory. Some of these improvements will be transparent. Go 1.24 will have a completely new map Implementation under the hood which is more effective on modern processors. And we prototy new waste collection algorithms designed around the capacities and constraints of modern equipment.
Some improvements will take the form of new APIs and tools so that Go developers can better take advantage of modern equipment. We examine how to take charge of the latest vector and matrix material instructions, and several ways whose applications can create in the locality CPU and memory. A basic principle guiding our efforts is Composable optimization: The impact of an optimization on a code base must be as localized as possible, ensuring that the ease of development in the rest of the code base is not compromised.
We continue to make sure that the standard GO library is sure by default and safe by design. This includes the current efforts to integrate integrated native management for FIPS certified cryptography, so that Fips Crypto is just a flag turns for the applications that need it. In addition, we evolve the standard library packages of Go where we can and, depending on the example of math/rand/v2Considering where new APIs can considerably improve the ease of writing of the CODE COTE and secure.
We are working to improve GO better for AI – and AI better for Go – improving Go's capabilities in AI infrastructure, applications and assistance to developers. Go is a great language to build production systems, and we want it to be a great language to Construction production IA systemsAlso. Go's reliability as a language for cloud infrastructure has also made a natural choice for LLM infrastructure. For AI applications, we will continue to build first -class support for GO in popular AI SDKs, including Langchaingo and Genkit. And from its beginning, go to improve the end-to-end software engineering process, so naturally, we plan to bring the latest AI tools and techniques to reduce the work of developers, leaving more time for funny things, as really programming!
THANKS
All this is only possible because of incredible contributors and the prosperous community. Fifteen years ago, we could only dream of the success that GO became and the community that developed around Go. Thank you to everyone who played a role, large and small. We wish you all the best in the coming year.
Credits: Austin Clements, for the Go team
Photo by Jason Leung on UNCLASH
This article is available on
