Let’s try to keep this topic focused on these first impressions and actionable ways we can improve them… and avoid an OS war. I looked at splitting some posts out but most of the incendiary comments are interwoven with on-topic content so I’m letting things stand for now.
7 Likes
Looks like at most 5 MB uncompressed, i.e. about a 1% increase in storage requirements.
(I just checked this based on the files required by a previous attempt at this, RFC: Add cygwin mintty to windows binary distribution. by twadleigh · Pull Request #12879 · JuliaLang/julia · GitHub, which installs the Cygwin mintty. It could probably be cut down further, e.g I’m guessing the Msys-based mintty — the same thing that “git bash” uses — is smaller. mintty.exe itself is less than 1 MB, the rest is system libraries.)
3 Likes
Well, Git for Windows compressed takes 44 Mb. And I’ve seen these guys expanding to hundreds of Mb (they pack Msys). That’s why I made that comment.
1 Like
Thanks to all who have responded. I appreciate the effort and time taken. Also thanks go to the ones who warmly welcomed me. In fact, may you all be well at this very moment.
→ Stefan Karpinski:
Yes, I think that (bundle Julia with a better terminal on Windows? #7267) was at least one discussion I saw where in the end no solution was given, as I was looking around for information regarding the REPL on Windows.
→ kevbonham:
To be honest, I don’t see how showing up as a new user and providing critique should be considered a problem. Must I have been here for a long time for my words to count? If I am new and tell you that 1 + 1 = 2, then does it not count? Or if I tell you that there is a mathematical set of such and such, then it does not count because I am new? (Maybe one would try to retort saying that these examples concern necessary truths or something of the sort and the main issues of the OP not, but one gets the point I try to make and the point is valid.)
I also don’t see why I must necessarily, if the constructive critique is good (as you seem to admit), immediately give concrete suggestions. I have stated at least one existing problem, a problem which has been admitted by the people in the issue linked to by Stefan, and I have then simply reminded the community, especially as a new user, that it needs solving. A concrete suggestion is actually already, without even mentioning it, implied.
You can probably expect that I, having been granted to be tech-savvy, have not solved the matter myself (and there exist people out there who are more tech-savvy and who would also find it a problem, since cmd.exe is simply not so good compared to bash or so, and it would be a problem to program something for this issue and consider all the pitfalls and horrible details that go with it), so what would befall others who are less tech-savvy? I can almost assure you that anyone, who will use the out-of-the-box Julia REPL on Windows, will be disappointed. That memory, that first-time experience with the Julia REPL, will always be retained. Because an REPL is a good way to quickly get familiar with the language, it is important that it is taken care of.
(I was actually thinking of perhaps trying harder to solve the problem myself or somehow help the community to solve it. In fact, I was willing to put some serious effort into it, but I suddenly for some reason don’t feel like it anymore. I wonder what that reason is.)
In any case, yes, I looked for solutions, as I implied in my OP.
I don’t trust Atom. Juno is based on Atom. I don’t trust Juno. Why do I not trust Atom? When I looked for information to refresh my memory about Atom, I was reminded of the fact that it used to take some sort of user information without the user’s consent (and I think there was at least one more reason). In any case, any piece of software that is known for such a thing is in my book already immediately to be reasonably doubted. Maybe, as far as I know, there is no problem with Juno, but you cannot simply and correctly dismiss my knowledge of Atom on which it is based nor the conclusion following from it, or am I a fool in this case?
As for RStudio and the so-called “official” Julia IDE, I did use words such as “so to speak” and “whether or not official” and “vim-like”. I think it should be clear what I mean (but perhaps I should have been clearer).
→ zhanglye:
Thank you (and also for the suggestions).
Even if RStudio has been existing for a long time, I have faith in the makers, maintainers and contributing community members of Julia, for look at what has been accomplished so far in what seems to be a reasonably short time. In fact, those two Julia IDEs mentioned look, at least if I check them quickly at a glance, pretty good, and they did not even exist as long as RStudio (or did they?). Am I then a fool to think that the people of Julia are very capable?
In fact, let me say this: am I a fool to think that, considering that Julia may be in the future a good alternative for R (due to its speed, and other good reasons), there should not be something like a standard Julia IDE, so that those RStudio people can have a pretty comfortable switch at least in terms of tools, IDEs and the like? In fact, anyone who would come to Julia would have a much easier transition, or a very smooth experience at least. I suppose it would more easily solve the REPL issue, while we’re at it.
Then let us think what could have been, if those two IDEs or at least one of them was still being maintained and improved up to this day. Perhaps by now you would have taken many, or at least would have interested (more greatly) many, who use both R and RStudio. Is this not something the Julia community would want? To get more users with all implied benefits? Then this would be part of the deal.
(more generally spoken) I do use Vim, by the way, but for R I use RStudio, simply because it is the most obvious choice to get going with it.
→ crstnbr:
When I tried cmdr, a virus was immediately detected and removed. Therefore I don’t trust it.
Let it be known, by the way, that my intent in creating the first post of this thread was not one of evil, nor is my reply in this post. I also have a kind of desire that Julia becomes greater and that its ecosystem becomes greater.
And, honestly, I was disappointed to read certain parts of some of the replies. I shall for the moment not point them out.
Abstract suggestions are great, but it helps to have a solid, actionable concrete suggestion so that people have an idea of what to physically type into a keyboard to solve the problem. But I’m sure you’d agree that this is obvious.
Whether or not you are a new user has no relevance to whether this is an issue. We also did not forget this is an issue, but there are many issues which need addressing, and only a limited number of people to address them. Feel free to contribute a PR to fix it.
They can read the docs, or wait until people have bandwidth to fix the problem. People aren’t eternally tech-illiterate, although some people seem to believe so.
Sorry, but to be quite frank, this makes you sound like an asshole. If you have the skills to fix it, then instead of complaining, fix the damned thing already. Whining helps no one, and as most of us are not paid to fix these issues, we have even less reason to help someone who can only complain and use passive-aggressive language to try to motivate people.
Yet another reason to not favor a single IDE! I use vim, which does no data collection. And since there is no single Julia IDE, I am free to use vim instead of Atom/Juno.
Please don’t be an ass.
If I was new to Julia and coming from R, I would undoubtedly be using vim instead of the RStudio GUI, so this would not be a “pretty comfortable switch” if now the only decent IDE was again a (in my opinion, no offense to the kind people behind Atom) very bloated, privacy-invading GUI IDE?
Are you implying that Juno is unmaintained? Or am I missing the point?
IMO: Either you downloaded garbage, or your AV is garbage. You’re on Windows, so either option is likely.
I would prefer if you did so we can either improve our future responses, or tell you that you misinterpreted our intentions.
2 Likes
Thanks @jpsamaroo for typing up all the things I was thinking.
No one thinks that, I assure you. And my response, and any of the other responses you might take to be critical are also not written with ill intent. Please don’t take it personally.
I’ll just note that when you came here with a number of complaints about Julia and its ecosystem. This is fine, constructive criticism is ok. But I and others complained and offered constructive criticism about the way you went about it. You could perhaps view that in the same light as you intended your initial post.
2 Likes
I also use Git for Windows and if the default Julia REPL used that (on Windows), it would be awesome 
Currently, I use four different terminals on Windows: 1. Julia REPL, 2. Git bash, 3. Node.js and 4. Docker.
If Julia used Git for Windows, I could probably figure out how to get the others to also work with it and I’d be down to one 
Edit: One reason I haven’t brought this up before is that I am unable to help implement this and I understand there are other priorities so I can wait.
Edit^2: Maybe it would be so awesome, but maybe not. I’m not sure, but checking it out and examining pros / cons would probably be a good idea if time allows 
I did confirm Git bash supports unicode. Feels like a new universe just opened up 
4 Likes
Yes, I would say that this is not a reasonable expectation. Also, even if there was a “standard” Julia IDE, it does not follow that it would look/work like RStudio, so I don’t quite see how it would necessarily make RStudio users “comfortable” (not that it’s anyone’s explicit goal AFAIK).
Being fixated on a “standard” IDE (which, BTW, RStudio isn’t, really — it is just one commercial alternative with a free tier), you also consistently ignore the fact that there are two existing GUI-based IDE solutions which are perfectly fine: Juno and VS Code.
Frankly, I am not quite sure what those benefits would be.
(Please note that many people try this kind of argument here: do what I want, and Julia will attract droves of users, which should somehow motivate you to work towards my goals for free; conversely, if you don’t, it will “not be taken seriously”. These arguments are usually delightfully ignored).
In any case, the bottom line is pretty clear: if anyone is interested in writing a tool like RStudio for Julia, then it will be available, otherwise it won’t. Arguing for it won’t make it happen, working on it may. But given existing alternatives which serve the same purpose rather well, I would not say it is very likely. You can of, course, develop or fund such a tool yourself.
9 Likes
@FRgs I think that evolution needs some time and we could hope that Julia ecosystem (and Julia community) will be more mature in the future. 
PS. And we (personally) could evolve too!
i have noticed over and over again that every now and then there will always be some new user complaining about an issue that has already been discussed a million times elsewhere, especially the IDE issue
you could start a new topic and announce to the world “wouldn’t it be great for mathematics if we could solve the continuum hypothesis or the Riemann hypothesis, because then so many more people would be more confident about X”
that’s not very productive, is it? the more productive thing would be, “trying Y specifically might help towards making Z feasible”
this would be much more useful towards an actual outcome, otherwise it is quite pointless to start this kind of discussion again (just search the discussions)
(not trying to criticize the OP, only trying to point out a general phenomenon)
having these questions under FAQ would help
3 Likes
I think that as a community we need to be better at taking criticism and just letting it be. Not every complaint or critique needs to be countered or refuted. Often the answer is: yes, that could be better and we will be working on it, thank you for the feedback.
65 Likes
Generally yes, but in this case the complaint is
and there is nothing here to address or improve. There is no criticism of Julia here, just a misguided impression that RStudio is somehow “standard” for R (it isn’t), and an arbitrary claim about not “trusting” Atom/Juno (both of which open source).
It is quite hard to just dismiss topics like this out of hand, because it is difficult to decide if the OP is simply misinformed and can be helped easily (this happens a lot), or not.
3 Likes
Do you mean Git bash for Windows? I cannot get e.g. \alpha to display properly.
Thanks for all the responses.
The first point is probably sufficiently clear, so it probably does not require anymore say from me for the moment.
Secondly, the more users you have, the more likely that there will be more contribution. (I thought that this part about implied benefits was obviously clear, and it may be that direct active contribution to Julia or its ecosystem in the sense of making certain products or tools or so would not have to be the only thing). Since RStudio is fairly common for R (forget about the nitpicking of words (such as “standard”, “official”, etc.), you should know already what I mean), strategically it would seem that, if you have such a thing for Julia, you will more easily attract R users using the fairly common RStudio (or other ones not even using R or RStudio). Perhaps there is something wrong with this argument? Maybe, as far as we know, but at the moment it looks fine.
Thirdly, the Julia people and the community, talented as they may be, are not the only people using their resources, doing things whether paid or for free. If I had at the moment sufficient resources, I would have liked to solve the Windows REPL issue. I was considering it: perhaps not do a certain other thing and invest in solving this REPL issue, but as soon as I read certain parts of certain posts, that desire was gone. (In fact, I was even already brainstorming in my mind about how to solve it.) Maybe it’s a problem on my part, maybe not, but even if so, lots of people are doing things, and so am I, and I probably do not have time to solve it. I have simply pointed out an issue or two, and perhaps I am not the one to solve it or them (where the first one, the REPL one, is more critical.) The matter has been pointed out, it has already been agreed with (at least in the case of the Windows REPL), and perhaps people need to be reminded and driven extra by an outsider who simply, for example, looks at that REPL and tells them like it is without dancing around the subject and what possible consequences there might be, and so on. In fact, is there now not suddenly discussion going on about the REPL in this thread, and does it not look as though it is going to be finally solved (by perhaps a good, simple and effective solution)?
–kevbonham:
A part of my background is philosophical in nature, and so I often enough had to deal with people who are, simply put truthfully without wanting to be offensive, horrible fools or people seeking out evil (not that I call anyone that in here) (and you might even wonder how they had ever gotten certain degrees or how they have been allowed to even hold certain high positions), who might not understand what is meant, or who twist something (whether with good or ill will), or who leave out important information from your words, or who do not care for truth, or who simply like to stab at people, or who somehow engage in fallacy, having no training at all in logic or argumentation, or no inclination towards such a thing, or even if they had it… So, whether good or bad, I am almost always or very often in a kind of critical mode where I try to check every word, every argument (if there even is one), and so on, with little or no liking for error (including my own, if I have erred to my knowledge). If there are errors in my writing in this thead, by the way, (if so, then in here at least most likely in language use,) I suspect it is at least because of a kind of battle fatigue (having lasted for years), heh heh heh.
–jpsamaroo:
I forgive you.
–Liso & StefanKarpinsky:
You have communicated wisely, humbly and truthfully. For this you may be recognized and praised.
1 Like
Regarding the REPL, there has actually been a lot of activity since you posted:
- https://github.com/JuliaLang/julia/issues/7267
- Fix color support on Windows via VT sequences · Issue #31491 · JuliaLang/julia · GitHub
- Changes to support color VT sequences on windows by musm · Pull Request #2 · JuliaLang/libuv · GitHub
It’s remarkable how complex and involved seemingly simple things like terminals, Unicode and fonts are. The problems are unfortunately due to various projects that are external—Windows and libuv (the portable I/O library that Node, Julia and various other languages use).
1 Like
there is a 
emoji in Discourse, but now I wish it was more graphic
1 Like