Julia 1.0 in 2017?


#1

Hello,

just received yesterday the Julia Computing Newsletter where it states that Julia 1.0 would/could be released in 2017 Q2 or Q3. Well, that is good news indeed if it’s the case! I guess I expected 1.0 to be out in 2 or even 3 years. Are we that close to a 1.0 release?

I ask because this would imply that I should begin to promote the language a bit more in my entourage, where lots of people are interested but are waiting for something more “stable” in terms of functions names, etc…

N.B. I know that 1.0 will be there “when it’s ready”. I’m merely wondering if it’s closer to a couple of months or closer to 2-3 years (as I expected).


#2

People have been saying 1.0 is coming out this year since last year.

Either way, it seems better to wait until 1.0 is out to tell your friends who want a stable release.


#3

Yes, of course. Some are more curious than others in term of early adoption.

I’m preparing a small presentation of the language in a couple of weeks for my team. I would have been more than happy to tell them that 1.0 is around the corner. I didn’t knew we were that close (in terms of time and certainly not im terms of work from the developer).

Cheers,


#4

I would be slightly cautious still. Julia v1.0 means that the Base language itself will have reached a stable 1.0. That does not necessarily mean that the whole package ecosystem will be setup with their own stable 1.0. I do see most major packages having a stable release by something like December (judging by how previous transitions went), but this is something to note if you’re saying “Julia will have a stable release” to people who don’t like being early adopters and plan to use a lot of packages.


#5

Considering that 0.6 is not out yet, I wouldn’t be surprised if 1.0 slipped to 2018.


#6

Yes indeed, that’s a good caveat to add to the affirmation.

But still, December 2017 or early 2018 is quite near and not a deal-breaker for a certain amount of transition towards Julia (mostly from MATLAB). I think people using Python on the team might stay with it.

Yes, that what I thought initially. That’s why the Julia Computing newsletter surprised me a little bit.


#7

The “rumor on the streets” is that v1.0 is supposed to be released this year because of the Moore Foundation grant, and “they” want to release it at JuliaCon. It’s a good narrative and maybe that was the original intention, but I too don’t know about the plausibility given the v0.6 delay. I would like to hear an update if possible.

If anything, most big organizations should plan to have a stable release for all/most associated packages just before Julia v1.0 so that way it has a solid foundation, and it would be good to start pruning projects / expectations.


#8

Even though I think that a stable release would be great, I am concerned that while the core team is hammering out “big” ideas, comparatively less attention is left to minor issues which are nonbreaking and matter little for performance (“language warts”). So I would not mind waiting a bit more for v1.0 if that means that more of these things are closed (I am not saying that it is up to the core team to do this, just that it takes time for any language to gather contributors).


#9

In my opinion it should not matter too much if Julia 1.0 happens within 2017 or 2018. The delta between 0.6 and 1.0 will likely not be to large so that an adoption during 0.6 would certainly make sense. I have been using Julia from 0.1-0.5 and the the amount of time porting to a new release was pretty small each time.


#10

I hope not. I get the point of waiting for every last wart to be fixed but the announcement in JuliaCon 2016 said 1.0 by the next JuliaCon. It would be a PR victory for that to be true. No one expects work on Julia to stop at 1.0, and the expectation was that 0.6 is 1.0 alpha.


#11

Let me put it like this: I’m more interested in stability and usability than in a revision number.

Currently i see too many ‘breaking’ or ‘decision’ tags in milestone 1.0 to be convinced that 1.0 is THE stable language and environment.

For myself i don’t see a problem, but if i’d try to sell julia as progamming solution within my employer, i already can visualise the mainstream of email replies…

But to end on a more positive note: Everyone asking for 1.0-when -> ask yourself, what you can contribute.


#12

Considering that 0.6 is not out yet, I wouldn’t be surprised if 1.0 slipped to 2018.

On the other hand, 0.6 is only a month behind schedule, and many of the major features we wanted in 1.0 are already in 0.6. In my opinion, the schedule for 1.0 has not really slipped – we’re still on track for this summer. I also think this is a fine time to start convincing people to look at Julia, while making sure they know a few more big changes are coming before 1.0, so if they really don’t want to have to update their code, they should wait until it’s actually released. But if they want to get ahead of the curve a bit, now’s the time.


#13

I’ve settled somewhere in-between: I convinced someone that wanted to leave MATLAB to have a look at it and another, a post-doc student, who wanted to learn another language, on top of MATLAB and Python.

For the employees of our team, I’ll wait for 0.6 and list the big changes between 0.6 and 1.0 and will prepare a nice presentation to convince them :wink:

You have to know that we are casual programmers and have a leg both in academic and private world. We program to obtain results and not to distribute software. In this case, I think the adoption is easier, as opposed to a firm that would need to support software (at least, in my opinion!).