“We are working on getting Sublime Text 3.0 final ready to launch”
forum.sublimetext.com
forum.sublimetext.com
After reading through the thread, the developer's points seem to be rational, coherent and well-delivered. I can understand some of the frustration from users, but not the outright hostility towards the end of the thread. They only have so much bandwidth, and the responses feel very "patient" to me.
To be honest, I've been using Sublime Text since before work started on version 3 and I don't really understand why there is such a fixation on the update cycle. Unless I'm mistaken it's only two full time developers, and historically it's been only one. The dev releases feel very stable to me, though of course I can't claim my workflow is the same as everyone else's.
I can't think of any particular features the text editor feels like it's missing. They seem to be nailing it, in my opinion. I ascribe to the philosophy that perfection occurs when there is nothing left to remove, not nothing left to add - they are relatively slow with adding new features, but the editor is blazingly fast and consistently delivers the experience I want in an editor.
For what it's worth, my opinion here is colored by the fact that I don't really think about the development cycle of most of the software I use. I couldn't tell you what major point release I'm on for 1Password. I think it's 6 on macOS, but I'm not confident. I just update it when it gives me an alert. Similarly, I don't really know what version I'm using for JetBrains PyCharm or CLion, and I've never followed the development roadmap. All I typically know is that I'm on the most recent version for the desktop software I use, and if something seems buggy or odd I check the changelog and then maybe look into it.
I'd be genuinely interested in hearing why people seem to be disappointed with Sublime Text, and what features they wish it supported.
Medium-term, I think there's a risk of Sublime falling away if it does not become a good client for the nascent "Language Server" ecosystem.
There are a few attempts at doing a client as a normal 3rd party package, but to be of the slickness that Sublime has established as its differentiator, I suspect it needs to added in-house.
From the Readme, here is a blog post[1] explaining what Language Server is.
[1] https://code.visualstudio.com/blogs/2016/06/27/common-langua...
This page says that SL3 already has good support for language server through a plugin.
Sourcegraph, who operate that site, took a run at doing a LS Client package for Sublime and it's been effectively abandoned for over a year.
They'd do a few regular beta releases, then just randomly _complete_ radio silence for months, a year or so - then, bam, another release, some "I'm not dead" activity on the forum from jon, then randomly silent again - maybe for a month, maybe a year - maybe forever.
The editor was good - although not free of bugs - but that complete lack of communication - did I just buy dead software? will there _ever_ be another release?
That makes it hard to have confidence that the software will continue to be supported - and hard to build a community around. The plugin ecosystem has definitely suffered a lot because of this stop-start cycle.
Bear in mind that the bugs that Sublime _does_ have often hit plugin authors hardest and they're the ones that need docs & help from the original developer, who's often been completely AWOL for months or years at a time.
Documentation is rare, often wrong, and the best you can do is try to find another plugin and dig through the source for that to determine how to implement something.
Or you can ask a question and... radio silence for months.
So I left.
I went to VSCode and Atom, both of which are more active and (total blessing for plugin developers) have source code available for viewing.
I really like Sublime. I think it's a great editor. I don't use it because I have better things to spend my time on than hoping a single dev gets the bandwidth to answer my question, or that I can hunt down some obscure plugin that happens to have the correct 'magic' that's not documented anywhere.
It can't even print. I'm a paying user but that one still blows my mind.
This final v3 might involve a licence upgrade but if there's one dev team I happily support it's these guys, what an awesome product.
Same experience. OS X, lots of plugins (Python, JS, Go, JSON, etc).
I'm not sure what to say about the belief some people have that 'we need releases twice a month' for a stable, high-performance product. I don't want or need product releases twice a month unless there's a good reason.
I feel like the regular releases crowd is mostly cargo cult driven by a desire to emulate the big web companies who are constantly - and often infuriatingly - tweaking their products for whatever reason.
From what I can see, this desire/need for constant tweaks arises mostly because web companies tend to ship very buggy products, and because there seems to be this belief that a company is going out of business if their web page that doesn't change often enough.
I love it. I could have stayed with ST2 for free but I chose to pay so that the devs could get the $.
I love sublime.
Very relatedly, I kind of wish TeX Live would release less frequently -- the 4-GB-of-crap song and dance is a pain, even once a year.
It shouldn't be too hard to combine https://github.com/thomasWeise/docker-texlive and a linode server or something into a LaTeX CaaS, either, but I can't say I've ever done so.
I don't feel bad giving them my money because I use Sublime for 8h a day, and given the amount of time between major releases (5+ years?) the cost per year is ridiculously low.
I use Sublime Text for personal projects I make no money from so obviously I don't want to spend out much money when I can get other excellent editors for free.
Genuine question: does this really make a difference in your day to day usage? I feel like once I've learnt the ~20% of features that are useful to me on a regular basis everything on top is marginal gains. Im really not that bothered by loads of new features.
I really really do not want to spend time to figure out on my own how to integrate this or do that (unless I need to) and the myriad of plugins/packages available to Emacs and Atom is something that is insanely helpful if you need to do something quickly.
I don't really need any updates from sublime. It does everything I need it to, and I've never felt a lack of a feature. I'm looking at the new and highly touted "hot exit" feature, and I'm thinking "but sublime prompts you to save when closing the app, and honestly I haven't lost work due to forgetting to save in several years."
> lots of communication
Again, I'm not sure why I'd want communication from the creators of sublime.
> great plugin ecosystem
Any plugin I've wanted (linters, syntax files, git integration...) has been installable from a dropdown menu in sublime. Maybe if my plugin needs were more esoteric, I'd appreciate vscode more?
> open source
I dig that, philosophically. Although if that was a priority for me, I'd go with Atom, instead of something published by microsoft.
Again, I'm open to the idea that vscode could be better than sublime - but maybe it's only better to people who need more obscure plugins, or who really care about open source?
but even if it didn't do anything more than Sublime, it's free.
But the most important thing is alive ecosystem - a lot of Sublime plugins are either dead or in maintenance mode, developers lost interest and moved to open source alternatives like Atom and VSCode.
P.S. Even though I moved to VSCode, I still love and respect Sublime.
I frequently point to VSCode as an example of a well-run open source project.
I will buy it again for the full price as a "Thank You!" to the developers. I have been using Sublime Text daily since 5-ish years and it has always been amazing. It's easily one of the best 50-60$ I have ever spent.
- You can easily stage all files, each file, or just specific hunks of specific files.
- You can stash and pop stashes.
- You can amend a commit.
- You can do all of that from the keyboard, without touching the mouse.
I imagine you can do many of those things with VSCode, but it's so effortless in Sublime. I haven't learned the same workflow in VSC yet, and since I do all of those things frequently, I'm really attached to SublimeGit.
One is C++, the other are hogging memory like crazy with Electron and a whole DOM.
If they're frustrated with Sublime due to not being FOSS, nothing I guess. But again, those people would have largely switched over already.
- sublime allows projects to have multiple different folders of files, vs code forces you to use a single folder
- large file support is unparalleled (excluding less etc)
- reliability, in nearly a decade(?) of using sublime I've never lost edits in a file or corrupted one
- it's general purpose so I can open anything, whereas vs code/atom seem more IDE-like (I know they can open anything, it's just my impression)
#2 was addressed in the latest update, and works well in my experience (open a 130M text file and jump to end in <5s).
#1 is available in the current preview build.
(I agree in general about Sublime performance vs Electron -- Sublime is way more responsive.)
Atom is a slow, hog, non-starter.
VS Code is almost comfortable to work with, and has more plugin polish, but it's in this "slow enough to be annoying and will never be fixed as long as its based on Electron" plateau for me too.
One huge irk I have with Electron is the amount of bloat it gives to even the most well written programs. With the VS Code example from before, no matter how well it performs, I always feel like I'm eating a stuffed crust pizza when I use it. Electron gives it much more cruft than needed.
I am currently debating on whether or not to switch back to Sublime since I'm getting sick of people using Electron as an easy way of building "quality" software.
But in fact, I use both. I always use 2-3 different editors. Mainly because using 2 monitors is easier to switch between apps than between tabs/windows of a app! So I use sublime to code and VS Code to visually something else. Or to write random stuff.
I have the same setup before with TextMate + Sublime.
I am using the dev build for my production work and it runs like a Swiss watch. The title of OP and the forum post is essentially a troll to instigate the developer(s) to come out and give a more definitive answer.
I wonder if they could look at staggered pricing system; where you can get:
N-1 release at £X; N-2 release at £(X-0.1*X) etc.
Where N is devel release channel.
That way people in N-2 will know what is coming and approx when.
People who want 'clarity' for the sake of it can pay less and STFU?
Dev -> Beta
Beta -> Final
Final -> Legacy
The current beta versions are rock solid. If you have been sticking with the previous stable version, you have been missing a lot for no reason.
IMHO this is the same mistake the Gnome team is making recently: Calling the finished version "final" or "stable". "Final" means to me "doesn't crash and has no bugs (as far as possible)". It doesn't mean that it receives no more updates. On the contrary, if it gets no more updates I would call it "legacy" or "abandoned".
We are definitely going to continue development of 3.0, so it won't become legacy any time soon. And yes, we do realize the beta period stretched on a tad too long. ;-)
Yeah, I guess the problem is now that people are not sure which version is supposed to be used for daily work, which version is still under development, etc..
But it is great to hear that development is continuing!
$70 dollars or so every 4-5 years?
For farmers yes. Not for developers in those countries -- which can make $30,000 - $50,000 in the international market.
http://www.urbandictionary.com/define.php?term=hellban
Not entirely accurate on HN since, apparently, all banned users can continue to post, but their stuff shows up dead (thus, they are only invisible to people with "show dead" turned off) and most of them seem to be well aware they are banned.
What a terrible thing to say!
Not everything always needs growth.
What else could they really add in besides a version bump?
It could also be that they don't make as much as you think. After all, 70 USD per major release is not really that much. Assuming, a 150k per dev salary (and it appears there are two), means they have to sell 4200 licenses or so. Sublime 2 has been around for 5 years ago? So that means they need 20k licenses. This only takes care of the salary though and not the hosting expenses, office space etc.
And last time I tried, working with large files was a pita on vscode.
For a professional programmer it should be less money that they make in a day for a tool they use 365 days every year.
And whether there are free editors out there doesn't mean anything as to whether it's cheap or not. Vim and Emacs and 20 other free editors already existed when ST3 (or SlickEdit, or whatever) came out.
Yes Emacs, Vim, Notepad++, etc. have been around for years but they are/were no way near as simple to get to grips with or as smooth out-of-the-box. That is where Sublime Text really nailed it. It was simple and beautiful from the moment you ran it. It didn't need a tutorial to help you figure out how to exit ;)
Now we have Atom and VSCode (to name two) that offer that excellent experience out-of-the-box along with regular updates, built in Git integration, open source, etc.
1) emacs is crazily extensible but hard to use. Uses hacks to get the performance 2) vi(m) is extensible, fast but also hard to use 3) ST is easy to use, fast but not that extensible as emacs/atom 4)atom/vscode is extensible,easy to use but slow
Because the game you'll finish in a few weeks and forget about and the meal you'll shit in the same day, whereas the editor you'll use for years on end daily.
Now, if you find another editor (free or paid) is better and gives you just as good or even a better experience, by no means, use that.
But if one feels ST3 is better, or even just a little better, then complaining about it costing 1/40th or less of a month's salary as if its any large factor doesn't make sense.
I think usually their dev/release cycles work pretty well and, to me at least, ST3 has been very stable for my usecases.
It's actually the piece of software I use for 6-10 hours each day for years that I've had the fewer problems or complaints about. And I'm using the dev 3 versions...