I agree that 5 years is probably too much for rocket start. What I am trying to say is that 1.0 as we could see it asap could not be version which you like to show to your boss to convince her to invest money.
IMHO in ideal world we have 2 options:
Repeatedly explain that 1.0 means only language stability freeze and it will be version with unprepared package ecosystem.
enlarge 1.0alpha phase more than is expected (see name of this topic). So we could have package ecosystem and user experience prepared for thorough reviews.
My guess is that we will see some kind of compromise between 1 and 2. It is because some packages could be prepared sooner and some later.
But wiki says: AAA game publishers have received criticism for engaging in ethically-questionable business practices. This often centers around the microtransaction gambling system called lootboxes, which are not currently regulated as gambling in most countries.[14] This system has been regarded as predatory towards children.[15][16]
I really hope that AAA business model (not only because what was cited) Julia won’t follow.
From weekly to daily? (Is it too obvious that I am newbie? )
And BTW my actual Julia 0.7 version is Commit 7fb5d27* (2017-12-18 08:00 UTC)
People criticize it because they paid for a product and the business uses them as testers. But that’s very different. With open source software, early adopters are testers. There’s no way around it other than to be really lucky and release a bug-free software on the first try.
I would highly recommend not being on the nightlies if you’re not working on the language itself.
Slightly offtopic, but I’d love to hear about your bad experiences with Juno (in contrast to VSCode) so we can improve (maybe in the gitter chat).
On-topic: Echoing @ChrisRackauckas I’d really like to have a stable API as soon as possible, just so that package authors can update their packages for the plethora of breaking changes. After the feature freeze there will hopefully be enough time for the whole ecosystem to catch up in time for the big 1.0 release.
As a long-time user of R and Python, I think R and Python by themselves are not so great (especially R), but they prosper with great packages. R and Python without packages are just like library without books.
For Julia, I think we should not ignore the fact its first area is numerical computation, just like Fortran, and the main concern is performance. The user community of Fortran is not that big, but very sticky. If Julia can be stable as Fortran without compromising too much performance, I think a loyal user community is achievable.
On top of that, building a successful ecosystem of packages is a long-term goal. Before tidyverse and pandas were available, R and Python were not as popular back then. At that time, I was using CERN ROOT and SQL to handle data. I feel like that after the concept of big data and machine learning came out, suddenly everybody was talking about R and Python. The first R code I used is from christopher sims, a winner of Nobel Prize for Economics, and that code is really old fashioned and academic, there is no fancy plot or pipeline operation. So I guess it takes more time for Julia to gain popularity.
My own experience is that Atom is not as responsive as VS Code, I have not measured the time, but it is noticeably slow, especially when it comes to using package and plotting.
It is not only slow by itself, it slows down my other software like R (RStudio) and Fortran (Code blocks). I remember that one time it even broke down my connection with a Postgresql server. I once asked some of my colleagues to try Julia, and I gave them an example code to draw a simple sine function graph, but it took about 40 seconds, that really scared people off.
I am not a cs guy or a marketing guy, but I kind of notice that most successful languages have enterprise-supported IDEs, like RStudio for R, PyCharm for Python, VS for C/C++ and C#, etc. I do not understand why Julia use a community-supported IDE as default.
Folks, please consider whether further posts here will be useful or productive. The OP’s question was answered, and I moved one more specific question to its own thread.
But the answer to the other general-complaints-and-comments is almost inevitably “we know”. If there’s a specific thing that is bugging you but doesn’t quite rise to the level of filing an issue, consider the #gripes channel on Slack. There have been many good issues filtered through there.
I agree that a feature freeze is needed, not that it should be called 1.0. Putting a feature freeze in 0.7 will be sufficient: Julia developers do read the Julia time lines, developers not (yet) into Julia will not.
Julia does not have the market clout (as does Microsoft, people already know and need Windows) to easily survive the fall-out of a botched first real release. If first time users experience extreme slowness and lots of errors following a basic tutorial (just making some graphs) they will stop at this stage (esp. as they come in expecting to find a fast language) and not look back. Also, are we absolutely sure that the slowness of precompilation and first use can be solved without breaking changes?
Well, i have to say ‘ridiculous’ has -at least for me- the definition of ‘you read it and start laughing’. This could be formulated differently, but politness doesn’t apply here. It’s based on observation and experience.
That statement is made in a specific context (i.e. that something is a “showstopper”). Please do not pull things from the context. If you want to tell Kristoffer something than please send him a direct private message. You attack Kristoffer in public here and even mention his employer which certainly has nothing to do with this discussion.
← my feeling when I keep reading words like “ridiculous” on this forum.
I don’t think @ScottPJones has attacked anyone. He shared his (correct) view about the lack of politeness of the word “ridiculous”, and highlighted that these words shouldn’t be used by someone from inside JuliaComputing towards a user of the language.
At least for me, a non-native speaker, the word ridiculous is very aggressive and demotivating. If I was admin I would hide that post instead of Scott’s post. That would have saved all this discussion again.
Not at all. I was trying to gently remind him to follow the Julia Community Standards:
Be respectful and inclusive.
and
Ask yourself if a comment or statement might make someone feel unwelcomed or like an outsider.
@liso had pointed out something that a number of people have complained about, which might be made better by having the precompilation done as part of Pkg.update (when you expect updating might take some time), instead of hitting you when you first use the package (you might not have even known you were using it, because it was used by some package used it from another package that you did use…)
There is absolutely nothing “ridiculous” about what @liso said.
Your saying that I attacked Kristoffer is very insulting to me.
I think that the employees of JuliaComputing should hold themselves to a higher standard of behavior, first off, because poor behavior on their part has the potential to damage the reputation of the Julia community as a whole, given that the company was founded by the co-creators of the language, and secondly, because JuliaComputing has directly received a serious amount of funding from the Moore Foundation, a large grant from the Sloan Foundation, as well as more recently investments from General Catalyst and Founder Collective.
I know that when I worked for InterSystems, there were very strict policies on how employees (and consultants like me) behaved when dealing with customers (or potential customers), especially in public forums.
I think it matters in the sense that he is not an ordinary user. His biased opinion can interfere with the progress of the language: if every time a user comes to give feedback and gets back a word such as “ridiculous”, the language won’t improve and the user won’t come back. I am myself considering leaving Discourse after many attempts to make it more civilized.
)