Allan Odgaard's Latest Comments on TextMate 2
old.nabble.com
old.nabble.com
This way, he sticks to the letter of the free upgrade promise - one can get it for free - yet benefits from what seems to be a very profitable way to price a product, especially a highly anticipated one towards which there is much goodwill.
(This does not mean I have any expectations as to when 2.0 is coming out, or what 2.0 means, or anything like that... I love Textmate, and have purchased several licenses.)
I completely believe Allan when he makes the "house without windows" analogy because this matches my own experience. As he said: don't ask for feedback when you already know what has to be done. For a version 0.1 you want feedback because you don't know which features are considered most important and you want to get hype going, etc. For a version 2.0 this doesn't work, because the ONLY FEEDBACK you're going to get is the whiny "but why doesn't it do X yet?" and "I wish you'd add feature Y", or "This icon is misaligned" or "There's a typo in this menu bar". The users mean well, but you know about all these flaws already because they're staring you in the face EVERY SINGLE DAY.
Releasing a complete rewrite is very difficult and takes a lot of stamina. As long as Allan is making progress every day (and I have no reason to question this) he'll get there eventually.
Getting "some kind of version" out there in 6 months would be a terrible mistake.
Anyone who has been burnt out should recognize that as a perfect description of it. It actually made me very sad to read that, since it reminded me so forcefully of times when I was burnt out - and I'm worried I'm close to that point again.
- "my motivation and optimism is anything but constant"
- "which will make me think it is all hopeless"
- "these will just be a reflection of my mood that day, which is anything but constant"
Allan may or may not be burned out. It isn't an unfounded observation, though. His comments sound a lot like burnout to me.
Right now, I love TextMate. It's the best text editor out of the box, but the fact of the matter is: OSX is not my only development machine and the hassle of learning Vim keybinds is less painful than only using OSX for development.
What I fear is that if 2.0 ever did reach a releasable state it will no longer be a light weight application I would use to edit everything from plain text configuration files, code, or a simple todo list. It won't have that same light weight do everything feel to it that the current version has. It will feel like a Visual Studio or Emacs. Which are good for their main purposes, but would you use them to edit a fairly small and simple configuration file?
TextMate, in my opinion is still by far and away the best modern day text editor available on any platform. I do like vim, but being a native Mac OS X application, TextMate wins out the majority of time (just drag a file onto the dock icon or typing mate . when in the terminal). It's visually pleasing and fast to open, browse, navigate and move around in.
The only features I really would like to see added is split panes/windows and full screen mode. Nothing is more awesome than MacVim with full screen mode and 4-8 split panes (on a single 2560x1440 monitor).
It's a difficult decision to make, the choice between improving an existing code base that has grown warts against the temptation of starting fresh and fixing what went wrong the first time, under the guise of learning from past mistakes.
I guess the lesson is: Don't try to rewrite your codebase from scratch unless you understand that you will spend an order of magnitude more time on the rewrite, even factoring in the experiences you have gained from doing it once.
But sometimes ripping out the plumbing and throwing the bigger pieces back together works wonderfully.
I guess that's called refactoring, though. :)
The point I'm getting at is, who does The Big Rewrite? I can't ever see myself wanting to rewrite sections of code that are working perfectly fine Just Because.
How much of The Big Rewrite was Mozilla? Did they literally rewrite everything? I can't imagine so. I bet they kept a LOT of code. It was probably more along the lines of a rather daunting refactor.
I'm just wondering what "starting fresh" means. For me, it normally means fixing architectural decisions that have become a huge pain in the but I'm keeping a LOT of the code base.
In many environments, specific refactorings are pre-packaged, such that they know their preconditions for working properly, and work with almost mathematical precision. In these sorts of environments, with additional support like Unit Tests, you gradually morph from one structuring of the system into another. This can be a lot safer than "ripping out the plumbing and throwing the bigger pieces back together." (I've also done the rip/reassemble thing. It's often fun and yields good results too.)
Of course, doing that as a one-man team with a closed-source commercial product that you have promised to give away for free doesn't seem like the kind of thing that can get done.
It worked out ok for end users, but didn't lead to such a great outcome for Netscape.
A few points:
- His experience sounds like "Second System" syndrome.
- One-man show is quite risky psychology-wise unless you can motivate yourself endlessly
- Perhaps it's time to hire another developers and move on to a more leadership role. Coding might be fun, but the risk of getting burn out is high; one way to prevent this is to do non-coding tasks for a while and once you miss your ol' coding lifestyle perhaps that's a good time to scratch your own itch again with hobby project. Who knows, it might become your second product, expand your portfolio.
I forgot about all the nice things I liked about TM. defunkt's textmate.el gets you part of the way there but isn't enough. I've had to implement lots of small things and discover features I didn't know about Emacs (global-auto-revert-mode). Right now I'm working on selecting an enclosed region (ie. find the innermost text surrounded by a pair of quotes or parens/braces/brackets and select it), which is Cmd-Shift-B in TM.
A friend and co-worker uses TM and the little niceties have tempted me to switch back since I mainly use OS X these days, however I'm simply better off with Emacs. TM is great but unfortunately while TM2 is in the works TM users are left out in the cold. There aren't many things Emacs doesn't already do due to its nature (open source and tradition of sharing code) and the things it doesn't do I can write myself. So I can bring Emacs up to speed with TM but not the other way around. Open source is a giant win for developer tools.
I don't want to bash TM or disrespect all the hard work Allan has put and is putting into TM. TM has influenced just about every modern text editor on many platforms and we are all better off because of TM even if we never used it. But if you're not happy with TM there are viable alternatives.
We all know the drawbacks to systems with a SPOF and it's not a good trade-off for a development tool as critical as your text editor. Sorry Allan, I truly am, but as TM2 takes longer and longer it gives people a big reason to migrate to another editor. Once switched, it's going to be a tough sell to get people to switch back without any guarantee that such a thing will never happen again. Especially if they switch to an infinitely customizable editor such as vim or Emacs where it's possible to mimic TM's best features, even if it's not as easy as in TM.
If any TM users do try Emacs get github.com/defunkt/textmate.el and have a look at github.com/samsonjs/config/tree/master/emacs for TM-refugee goodies.
I just want few small things fixed in current version, e.g, asynchronous refresh of project directories (or at least better version of remate hack http://ciaranwal.sh/remate/ ).
Finally, I'd pay additional for a Linux version.
I bought TextMate about 2 years ago and couldn't be happier with it. I've extracted way more value than the ~$50 I spent on it and would be happy to plunk down another $50 for TM2 if there were compelling reasons.
It seems unreasonable to buy software solely because of what you expect out of it tomorrow. You should buy it for what you need today.
EDIT: I'll add that I don't think that this is a greedy proposition. While I know a lot of people are truly thrilled with TextMate, I wasn't. I didn't ask for a refund, because I understand what it's like to be an indy developer, and thought I'd gamble that Alan would ship the features I'd need to switch to TextMate over Emacs when the 2.0 release got out the door. That gamble didn't pay out, which I'm okay with; 2.0 never shipped, and I'm not mad that 1.0 failed to materialize with the features I needed. But I also don't think it's fair to say I shouldn't get the 2.0 release, when I almost certainly would have asked for a refund based purely on the 1.0 release. I doubt I'm the only one in this boat--especially given how many blog posts I've seen over the last several years as developers, frustrated with TextMate's lack of progress, have reverted to Vim/Emacs/BBEdit/what-have-you to get the pet features they want.
What's really unfortunate is that the handful of features that the majority of people are really itching for don't require the kind of Big Rewrite/second-system effect that's currently going on. It's frustrating.
There's nothing bad about ridicule. A product built under gaze of a select few is destined to be far removed from the needs of its users. TextMate was not born good, it improved because it was ridiculed.
That said, I'm no longer an interested party. I stopped using TM over a year ago after at least a year of using it and seeing where it was going.
Yup. Probably my least favorite part about writing software. And "project management" types still expect time estimates. :)
Also I was not aware that Textmate was the product of a single developer. (is it?) It's actually pretty motivating to consider!
Last time I checked (a few months ago) it was pretty slow and unstable, but I guess the project is worth watching if you're not on OS X.
Shipping is probably one of the hardest parts of software development. I hope Allan can get through it and understand that people are going to be critical of TM2 only because they love TM so much and want it to be awesome.
> glaring problems, like unthreaded project-wide search
I've seen a few threads now about how an "Updated UI" would be a welcome improvement from the 1.x branch; I'd agree with that.
I really wish there would be some competition to heat things up. I am not talking about emacs or vi, but another great visual editor for OS X. Maybe Open Source... or maybe a startup could do this. There should be quite some money to be earned.
Or is there something comparable to TextMate that I don't know about?
Sublime Text (www.sublimetext.com) could theoretically be ported over to Linux and Windows as it's GUI is all OpenGL and C interfacing with python. The author's original intent was that it be portable between platforms, but he seems to have pushed that idea to the side. He's no longer working on it full-time either. Which is a shame because it's quite good and comparable to TextMate in many ways.
TextMate is an extensible platform like Emacs, and so it is constantly evolving thanks to people's generosity in writing and giving away plugins.
I do a lot of Lisp work in Emacs but often keep TextMate open in parallel. Neat tool, and worth the money. If the author never brings out version 2, that is his call.
I'd go as far to say that he has every right to charge for TM2 when it comes out, even if he promised it would be a free upgrade.
EDIT: even though I've stopped using it, I use vim now :)
XCode's editor is OK, and replacing it with TM may take more effort than adding few features it's missing.
http://intype.info/download/download.php?intype-0.3.1.734.ex...