This is fantastic. I've been an addict of http://reddit.com/r/artisanvideos for quite a while, because there's something fascinating and enlightening about watching artisans perform their work (for example, Alexis Ohanian designing the Hipmunk bellhop logo - http://www.youtube.com/watch?v=CYEQpwduyPU).
It's especially interesting to note that there are few, if any, words in the entire essay that he didn't edit multiple times. A lot of times when you read something as well-written as PGs essays you get the feeling that the person writing them just sits down and hammers out brilliance. But in reality, at least in the case of PG - and I would venture a guess that most great authors are similar - greatness comes through sweat and repetition more than raw talent, until eventually that thing becomes second nature - or does it ever? I would kill to watch some of the great authors of all-time (a la Shakespeare, Hugo, Dickens etc.) write in real-time - it would be fascinating to know what their process was like.
> Everyone revises their work, many times over. But yes the great authors do as well.
You state that as if it's some universal truth but it's certainly not. We know that Asimov only ever wrote a first draft and a final draft and Heinlein wouldn't even create a first draft most of the time. Neither of them is a Nabokov or Hemingway but they certainly fall into the category of "everyone."
'I remember the players
have often mentioned it as an honour to Shakspeare, that in his writing (whatsoever he penned) he never blotted out a line. My answer hath been, "Would he had blotted a thousand," which they thought a malevolent speech. I had not told posterity this but for their ignorance who chose that circumstance to commend their friend by wherein he most faulted; and to justify mine own candour, for I loved the man, and do honour his memory on this side idolatry as
much as any.'
I say go ahead and compare. Shakespeare and Dickens were both considered highly-commercialized and not art in their time. We already see pg's incredible value as a writer.
And what does the format matter? If it did, we would disregard haikus as poetry just because they can fit into tweets.
For a roll-your-own version of this, you can use the 'script(1)' shell command, the timing argument, console editor of your choice, and 'scriptreply(1)'.
Script is a Linux utility which makes a typescript of a terminal session. Traditionally this was mostly of use for shell sessions which didn't involve full-screen (ncurses) utilities, as the cursor-movement characters would typically present a jumbled mash on your screen.
Using the '-t' (or '--timing') arguments sends timing between movements to stderr, or the specified file. With this, you can then replay a typescript file, with the timing information, to show what was actually presented on screen. An optional argument varies the replay speed.
I've made use of this to log and replay serial console activity (available on most servers through IPMI or related hands-off / lights-out management tools), which can be both instructive and helpful in documenting steps.
Further fun may be had by playing the output of scriptreplay through the phosphor(6x) X11 screensaver hack.
The interesting part about watching this isn't watching it being written. It's watching it being edited. What to leave on the cutting room floor is always the hardest part, and Stypi's "Doomed" feature does a great job of showing you how someone comes to the decision to self-edit.
Cool. However, I feel that watching it, doesn't really give the full effect, requires tons of patience too. Couldn't the editing data be visualized in some data science illustration for a real insight into what it represents?
Maybe like the final draft, overlayed with a heatmap of how much editing was done, then on clicking, provide a tree like comment history of each branching edit. This could show the anatomy of how different people form thoughts in written word. T his is an insight we could learn much from at a massive scale.
I would imagine Facebook or another hosted blogging platform could get this data with relative ease.
This functionality was in etherpad, which was also YC-funded and also demoed their software with a playback of a pg essay.
While etherpad lives on as an open source project the original company/team was acquired by google and this resulted in google docs getting real time cursor movement and edits shortly thereafter (google docs originally updated via polling and often had collisions).
I think it's a shame that google never implemented more of the core ideas from etherpad, such as playback and contributor text highlighting. It's still the best online collaborative tool I've ever used. I had experiences writing letters to politicians in real time with dozens of anonymous strangers who we gathered via twitter, and collaborative grant proposal writing sessions that I can only describe as thrilling. It's also been fun to use etherpad to collaboratively take notes during conferences, and truly fascinating to replay and see how largely anonymous editors quickly start taking on certain roles without any prior discussion--one person gets very interested in fixing spelling mistakes, someone else gets really good at finding relevant urls for background information and links to slides, someone else just barrels ahead with whatever the presenter is currently talking about knowing that the others will fill in the gaps. A lot of the innovation happened in subset hardier, but etherpad made it way more accessible and this useful.
Etherpad deserved all the attention that google wave was receiving at the time because it really delivered on the promise of a truely flexible tool for collaborating and communicating more effectively than email.
> I think it's a shame that google never implemented more of the core ideas from etherpad, such as playback and contributor text highlighting
> Etherpad deserved all the attention that google wave was receiving at the time
I find your post confusing because I thought that Google Wave did have playback and contributor text highlighting.
I suspect that the reason Docs never gained these features was that Wave was supposed to replace it, but AFAICT Wave never saw the kind of user adoption it needed (probably due to its sluggish performance) and was killed off.
This was my favorite part of this product, the highlighted text is "doomed" and will be altered or removed at some point in the future. It's great to see how much editing happens after the initial draft.
It represents "doomed" text that eventually gets removed from the final draft. You can toggle it by clicking "doomed" in the upper right. See http://paulgraham.com/stypi.html
This is fantastic, I thought I had some sort of dyslexia, because of my non-straight, back and forth type of writing :-P
Good to know I'm not alone. The more I read HN the more I find people, who are successful in their field and have the same behaviors as me. It's relaxing :-)
I'd really love to hear pg's comments of wether there is a piece of software, a tool, that could help him write the essays. Is there something that constantly comes to his mind when he writes, in form of "if only there was X so that I could Y".
Not really. Vi is all I need. I could probably think of things I'd like but don't realize I need. E.g. controlling an editor directly with my brain, so I didn't need to type. Hard to say about the narrower question of whether there is anything I'd want that's possible with existing technology.
Thank you Paul. I'm missing an ability to collapse / uncollapse parts of written text (meaningfully), so you could see only the plan, or summaries of each chapter, or paragraphs, etc. It would require a very good interface to be useful, but imagine the implications for consuming / learning when knowledge is presented in this way.
honestly been wanting to build this for myself forever
I guess I'd answer the question with a cui bono question. Why would someone of Paul Graham's standing need to fake this? What would be his incentive to do so? He has nothing to gain from faking it. And I can't think of anyone else who stands to gain, either.
Occam's Razor explanation is that he just thought it would be cool to show.
Without the link from his own site we can't tell who even claims to have done the typing. Someone who isn't PG may have all sorts of reasons to fake it. Gentle suggestion - if you think the answer to someone's comment is obvious then it's a good sign there's a miscommunication and you should consider other interpretations.
"Gentle suggestion - if you think the answer to someone's comment is obvious then it's a good sign there's a miscommunication and you should consider other interpretations."
To the contrary, I didn't think the answer to his comment was obvious. I don't claim to have the correct answer. And if that's the tone my comment conveyed, i.e., a dismissive one, then I probably conveyed myself improperly. That was not my intent.
Real time, as in a 1:1 ratio between the time it took him to write it and the time it takes for you to see it played back. Not as in live, where you see it as he's typing it.
I'm still waiting for XEP-0301 to become ubiquitous (which transmits key delays with text, and would be closer to what the GP thought he saw) - I really miss this feature in chat systems.
It's especially interesting to note that there are few, if any, words in the entire essay that he didn't edit multiple times. A lot of times when you read something as well-written as PGs essays you get the feeling that the person writing them just sits down and hammers out brilliance. But in reality, at least in the case of PG - and I would venture a guess that most great authors are similar - greatness comes through sweat and repetition more than raw talent, until eventually that thing becomes second nature - or does it ever? I would kill to watch some of the great authors of all-time (a la Shakespeare, Hugo, Dickens etc.) write in real-time - it would be fascinating to know what their process was like.