The vision here is noble, and I applaud what Chris has done with Light Table, but based on the discussion here, I think the Eve team is setting themselves up for disappointment.
1. By bringing investors on board, and promising Hacker News they're going to change the world, they're basically asking to be stressed out all the time. This is not conducive to innovation.
2. There's a fundamental tension between creating something usable by Joe in accounting and something that is cutting-edge from a technical perspective (highly concurrent etc). 90% of line-of-business programs use less than 10% of CPU, and approximately 0% need high concurrency. Joe in accounting is not going to build the next WhatsApp, and every moment you spend figuring out multi-core whatever, you're not thinking about how to help Joe make his Cash Flow Prediction Tool look slick.
3. From a marketing perspective, creating a new category of software is an enormously difficult undertaking. There are two related problems here:
A. You have to find potential customers. Because nothing like Eve exists, the market does not exist. I.e. the world's "creators" don't all go to the same conferences, visit the same websites, and hang out in the same IRC channels. Light Table at least was addressing an existing market, i.e. "programmers who use X/Y/Z", so you could actually find customers, and they could tell their programmer friends about it.
B. After you find potential customers, you have to educate them, i.e. explain the potential benefits and teach them how to use the darn thing. This requires a major investment from both you and from your interested customers. It is much, much easier to sell something that relieves the customer's pain than something that promises him "superpowers". 95% of "creators" have only vague and incorrect notions about programming, so you'll have a hell of a time explaining the product. Let's say you want to target Joe in accounting: What will the advertisement actually say? "Do you miss VB6?"
4. The fact that the LT team is giving up on Light Table when it's still the proverbial "couple of guys" is not a good sign. From a marketing perspective, Light Table had every advantage in the world: tons of publicity / viral videos, a clearly defined market, and an exciting product that got people talking. The fact that they're throwing in the towel tells me, well, they quit too early, and they'll quit on Eve when the going gets tough.
To be more specific, I think Chris's fears about the market for Light Table are totally unfounded. To wit:
* "The competition has millions of man-hours invested". Who cares? On a day-to-day basis I execute maybe 3% of the code paths in my editor. In my view, Light Table is a classic opportunity to offer 10% of the functionality of the bloated competition but really knock that 10% out of the park. Look at e.g. Pixelmator versus Photoshop or [toot toot] Wizard versus SPSS.
* "Programmers don't pay for software". I'm sorry, but this is a poor excuse, and if I were your drill sergeant, and this were an 80s movie, I'd be screaming in your face right now. Programmers are a wealthy and growing segment. They'll pay for stuff if it helps them get their job done (GitHub, TextMate, books, conferences, heck Visual Studio). Not only that, with the product's positioning, you have a great opportunity to sell to people who want to be programmers, the "ski pants" market, if you will. If you read the comments in this thread, you'll notice that your users are complaining about the Light Table's lack of polish and they clearly recognize the need to pay someone to provide that polish.
In sum, I think the LT team should have worked themselves into a crying, bleeding, starving mess to make Light Table a commercial success, and then pursued bigger ideas once they had the business experience under their belt. Even if it took a few years, they'd be in a much better position both financially and psychologically to tackle a Grand New World-Changing Product once they had the routine down from their Small Life-Changing Editor.
As an indie-app-developer-whatever myself, I would kill for the kind of publicity and market opportunity that Light Table had. I wish Chris & Co. all the best with Eve, and I truly hope they succeed. But based on what I'm reading, I don't think they don't deserve to.
Hey Evan, I have immense respect for your work, but this is surprisingly armchair quarterbackish.
> "The competition has millions of man-hours invested". Who cares? On a day-to-day basis I execute maybe 3% of the code paths in my editor
This is the Microsoft Office argument and the problem is that no one uses the same 3%. If they did, sure it'd be "dead simple", by that of course I mean take many, many years of work. Sublime which attempts to do far less than we wanted to took 5 years to get to today and I'm not sure it really is a commercial success capable of supporting 3+ people.
> "Programmers don't pay for software". I'm sorry, but this is a poor excuse
It's far from an excuse, it's the truth. It doesn't matter if developers as a class have a lot of money. They don't actually spend it. I don't know a single person who has ever personally paid for Visual Studio. Github is a service and the vast majority of their money comes from companies not individuals. How many people run stuff only on the heroku free tier?
Or here's a direct example: when we launched the kickstarter people \freaked\ out about us pricing it at $50. We had to drop it all the way to $15 and folks still complained endlessly about it. It is well known that short of a service targeted at companies, it is near impossible to sell developer tools. After all, it's a group of people who look at a piece of software and say "I can build that".
If there's some secret, we're missing it. There's an entire organization [0] that has grown up around trying to help dev focused companies figure this stuff out. It's led by James Lindenbaum (co-founder of Heroku, invested in us), he knows his stuff and he couldn't figure out a way to make this work out either.
You need to understand that people like to complain about stuff. If they complained about the $50 and then you dropped it to $15. They got exactly what they wanted. Why would they not just keep complaining to get it down another $5?
People pay for stuff that is useful to them, the problem with light table is people don't understand why it is useful to them until they think about it/look at it. You should have completely ignored every complaint related too money.
VS costs hundreds of dollars (going into the thousands for MSDN). And it is used by hundreds of thousands of programmers.
If product A results in a increase in productivity people will pay for it.
Microsoft is an irrational actor: they will sink billions of dollars into VS without regard for direct return in order to secure Windows as a platform. Using them as an example is useless. Moreover, VS is "purchased" almost exclusively as part of the entire Microsoft stack. When I was there a tiny, tiny fraction of revenue came from anyone actually just buying VS.
Why? It's probably one of the most popular editors for PHP/HTML/CSS/Javascript/Ruby/Python.
According to https://sublime.wbond.net/stats there are roughly 2.5M sublime text users. If even 10% of them paid for it, that would be 250,000. Assuming 10% is a little high, let's do 1% would would be 25,000.
Even if those numbers are wrong (which they may be I have no idea how he gets those numbers, and if everyone is using Package Control) there can very easily be hundreds of thousands of paid licenses.
That is certainly true, though I haven't seen many IDE-based consultancies. In our particular case, that wouldn't have been the right direction - none of us want to be "programming", much less doing it at other people's whim.
Ah yeah, Eve has a ton of interesting business opportunities around it, everything from on demand computation/hosting (heroku) to collaboration and versioning (github). Consulting or even just building products of our own that are much easier to maintain/faster to build is also an option.
I know Chris already responded to some of your concerns, but I'd like to weigh in briefly as well.
> By bringing investors on board, and promising Hacker News they're going to change the world
a16z constitutes the majority of our investment, and Chris Dixon at a16z is rather well equipped to help us tackle this problem. It has also given us access to a rather impressive network of very smart people who can help us. Yes, there is a trade-off in that our investors have a say in what we do, but given the small number of investors, and who they are individually, there is a low probability of micromanagement from them. Secondly, we never promised Hacker News that we're going to change the world; we said we have some ambitious ideas, and that we wanted to pursue them. Might they change the world? Sure. Might we also die trying? Absolutely. That's a risk inherent to... well, everything. Certainly a lot of things worth doing.
> The fact that the LT team is giving up on Light Table/I think the LT team should have worked themselves into a crying, bleeding, starving mess
The last time we had office hours with pg about LT, we didn't exactly leave with a warm fuzzy feeling about continuing work on the project. But we did, for another year, spending the remainder of the Kickstarter money and approaching the end of the runway, staying as lean as possible, and continuing to develop a product that pg, James Lindenbaum, et al, all regarded as a nonviable business. We may have to agree to disagree on how plausible it is to make an open source IDE a commercial success, but there was plenty of crying and bleeding before we came to work on Eve.
At that point, it became what Jamie aptly described as a matter of return on effort. There isn't much sensibility in spending countless hours to solve the wrong problem, and running out of money didn't strike us as the optimal way to continue working on a problem we genuinely cared about. Instead, once LT became open source, we forged ahead into an area that Chris Dixon called a "Vietnam" of software, and while Eve may be particularly difficult, it brought us new funds precisely so we don't have to give up.
I was waiting for LT to become a tool I can recommend to beginners. It's far from being ready, and it now appears they just gave up. If LT was too hard for them, what are they even talking about with this Eve thing?
It's not a matter of difficulty, it's a matter of return on effort. We could pour hundreds of man-years into catching up with emacs/vim but we would be solving the wrong problem. The reason that language tooling is so hard is because languages are not built for this kind of tooling.
Why would anyone want to catch up with emacs/vim? The original idea behind LT was visual programming, as it was shown in the Bret Victor's demo. That was the whole point, and I don't feel LB provides that yet. That's why I'm sad they gave up on it, and started a new project, which appears to be even more difficult.
Programs are written in text. If you want to edit programs, you have to be reasonable text editor. You need find-and-replace. You need to handle files being updated underneath you. You need syntax highlighting and incremental parsing. You need to be able to match error messages to locations in files (much harder than it sounds). You need to be able to open large files without dying. You need to scroll smoothly. You need to handle custom key bindings, different keyboard layouts (eg https://github.com/LightTable/LightTable/issues/620) and unicode editing.
You have to do all these things because if you don't the editor is not usable for editing programs. The vast majority of the code in LT is devoted to basic text editing. Just reaching the same level of polish as, say, gedit would take many more man-years of work. All of that work has to be done before we can spend time on actual innovation.
Or we could just make a language with a sane serialisation format. Eve already supports live coding, incremental compilation as you type, globally unique identifiers for every piece of code and time-travel debugging. We also have algorithms sketched out for concurrent editing, commit-as-you-type (ie the version control system stores every keystroke) and causal history for data (why/why-not provenance). We got to do all that in a few months because we didn't have to fuck around building a text editor.
Wow, what you described is not at all what I was looking for when I supported Light Table. I already have IPython IDE where I have all that text editing, and all other stuff I need to type in and run my code. I was looking for something completely different, not yet another text editor! I was looking for something visual, where I can see what my code is doing without me having to insert a bunch of print statement.
In fact, I found some of that here:
http://pythontutor.com/visualize.html#mode=edit
but I hoped that LT will do it even better.
I think their new project plans to fulfill Bret Victor's vision in a way LT cannot. Bret Victor's demo was meant as a starting point. He was really advocating tools that help us understand the systems we create as programmers. Not live editing. I think Eve is an attempt to give us the tools to build and reason about our systems.
>> "they clearly recognize the need to pay someone to provide that polish"
I don't see anyone offering to pay for polish. Can you point to these instances?
I've been active in the LT community for most of a year and don't recall _one_ offer to pay for any bug fix or feature. If anyone has such a request, please make it in LT's github issues or mailing list and I'll consider looking into it.
I think the (relative) success of RubyMotion as a paid product puts the myth that programmers won't pay for software to bed. I haven't bought RubyMotion yet because it doesn't meet my present needs, I would buy LightTable (for an RM order of magnitude price) because it does. What LT will probably not do though is sell in a way that pays back VC's looking for a go big or go home hit.
1. By bringing investors on board, and promising Hacker News they're going to change the world, they're basically asking to be stressed out all the time. This is not conducive to innovation.
2. There's a fundamental tension between creating something usable by Joe in accounting and something that is cutting-edge from a technical perspective (highly concurrent etc). 90% of line-of-business programs use less than 10% of CPU, and approximately 0% need high concurrency. Joe in accounting is not going to build the next WhatsApp, and every moment you spend figuring out multi-core whatever, you're not thinking about how to help Joe make his Cash Flow Prediction Tool look slick.
3. From a marketing perspective, creating a new category of software is an enormously difficult undertaking. There are two related problems here:
A. You have to find potential customers. Because nothing like Eve exists, the market does not exist. I.e. the world's "creators" don't all go to the same conferences, visit the same websites, and hang out in the same IRC channels. Light Table at least was addressing an existing market, i.e. "programmers who use X/Y/Z", so you could actually find customers, and they could tell their programmer friends about it.
B. After you find potential customers, you have to educate them, i.e. explain the potential benefits and teach them how to use the darn thing. This requires a major investment from both you and from your interested customers. It is much, much easier to sell something that relieves the customer's pain than something that promises him "superpowers". 95% of "creators" have only vague and incorrect notions about programming, so you'll have a hell of a time explaining the product. Let's say you want to target Joe in accounting: What will the advertisement actually say? "Do you miss VB6?"
4. The fact that the LT team is giving up on Light Table when it's still the proverbial "couple of guys" is not a good sign. From a marketing perspective, Light Table had every advantage in the world: tons of publicity / viral videos, a clearly defined market, and an exciting product that got people talking. The fact that they're throwing in the towel tells me, well, they quit too early, and they'll quit on Eve when the going gets tough.
To be more specific, I think Chris's fears about the market for Light Table are totally unfounded. To wit:
* "The competition has millions of man-hours invested". Who cares? On a day-to-day basis I execute maybe 3% of the code paths in my editor. In my view, Light Table is a classic opportunity to offer 10% of the functionality of the bloated competition but really knock that 10% out of the park. Look at e.g. Pixelmator versus Photoshop or [toot toot] Wizard versus SPSS.
* "Programmers don't pay for software". I'm sorry, but this is a poor excuse, and if I were your drill sergeant, and this were an 80s movie, I'd be screaming in your face right now. Programmers are a wealthy and growing segment. They'll pay for stuff if it helps them get their job done (GitHub, TextMate, books, conferences, heck Visual Studio). Not only that, with the product's positioning, you have a great opportunity to sell to people who want to be programmers, the "ski pants" market, if you will. If you read the comments in this thread, you'll notice that your users are complaining about the Light Table's lack of polish and they clearly recognize the need to pay someone to provide that polish.
In sum, I think the LT team should have worked themselves into a crying, bleeding, starving mess to make Light Table a commercial success, and then pursued bigger ideas once they had the business experience under their belt. Even if it took a few years, they'd be in a much better position both financially and psychologically to tackle a Grand New World-Changing Product once they had the routine down from their Small Life-Changing Editor.
As an indie-app-developer-whatever myself, I would kill for the kind of publicity and market opportunity that Light Table had. I wish Chris & Co. all the best with Eve, and I truly hope they succeed. But based on what I'm reading, I don't think they don't deserve to.