Select text to annotate, Click play in YouTube to begin
00:00:15
okay and it just gets better Jeff come on up on stage Jeff Rollerson will come back we're gonna have a conversation about tools and we have other panelists in this room so Dan where are you Dan come on up here Dana
00:00:29
and Jeff he's coming okay okay why are you in the back of the room okay no okay just sit in the outside one end yeah Brett couldn't make it but we're gonna have a little demo of his so I actually
00:00:42
didn't collude with Vint on this but I plan to follow on what we're gonna try to address Doug Doug had many fantasies
00:00:57
for NLS which we did not get into the demo there was actually some heated discussion about some of them but one of the things that was that he was pushing
00:01:09
on even back then was the idea that that well we've heard a lot about documents and people produce a document and then people link a document and they make links to other documents Doug wanted
00:01:27
wanted a kind of system where you could sit with your monitor or your display and your keyboard with other people and
00:01:38
and sort of resolve he wanted the system to be able to help you not just think things through but resolve arguments and
00:01:50
he had ideas about about highlighting text in different ways and and how the links would work so that you could use it sort of in facilitating the understanding to bring resolutions to
00:02:03
heated discussions and things like that and another idea that he had was was linking linking documents together in such a way that that the viewer would get a personalized view of what was
00:02:23
going on that is if I looked at it I would see something different than if when Vint looked at it and I actually implemented some facilities to do that nobody knows about him because they were totally
00:02:36
unusable and very confusing and and you had to be a extraordinarily skilled educated computer scientist to touch
00:02:49
type code to make them work because you actually wrote the programs to direct the searches and stuff like that so they never saw the light of day although they were some of the code wasn't a demo so
00:03:02
sort of playing on that we've got Dan and Jeff here and they've they've bent their pick trying to do these things broken broke their fix trying to do them
00:03:15
so what we're going to do here is have them each present what they've been working on and then I'm gonna try to I can take what they're saying and have a little discussion with them to to try to
00:03:28
map it back into what I call bootstrap speak if you think that H Lampe he had all you know hundreds of these wild ideas and he had abbreviations for all of them and I've always thought of
00:03:44
that as bootstrap speak so let's see who wants to go first dan okay go for it hi Doug understood deeply that we face extraordinary challenges not just
00:04:05
because the complexity and urgency of the problems we face are outstripping our capabilities but because knowledge and capability is increasingly specialized itself really hard problems think of the Fukushima nuclear incident
00:04:19
are increasingly only able to be solved by a smaller and smaller percentage of the overall population two concepts were key to his thinking one we've heard about today collective IQ the the
00:04:32
community's ability to deal with complex urgent problems but the other was the DKR for the dynamic knowledge repository which was a live living breathing rapidly evolving repository of all the
00:04:45
stuff accumulating moment to moment for a community think of Wikipedia today I'll tell you about another kind of DKR developed by a small group of noble individuals from different organizations
00:04:57
working together over the last decade so the problem is that right now there's no easy way to know the thinking of other people about the world around us there might be tens or thousands of people
00:05:10
that might have read a particular article or book and inevitably a number of them will know a considerable amount about that subject so that on this document or page or book or PDF that I'm
00:05:23
on right now how could I be informed by their perspective and how could I tap into such a thing without it being an immediate vector for spam and trolls and the worst everything online we do have a
00:05:36
thing now we call comments but comments exist on only a tiny percentage of content on the web considerably less than 10% comment widgets are put on pages and moderated by the people that own websites plus comments are below the
00:05:50
fold proprietary technology that only allows a single layer of contributions the system in one place is not the same as a system in another they don't work on PDFs or ebooks you can't be universally logged in you don't own the
00:06:02
comments you make and you can't search them or tag them or make private ones for purse notes and the fact that they exist on only a tiny percentage of content really renders them effectively useless what we need is something we can carry around
00:06:15
with us what if instead we could annotate things what if over every preprint on archive for instance you could see if you wanted layers of annotations from your scientific peers providing additional insights
00:06:28
what if overdraft legislation coming out of subcommittee you can see the annotations of citizen advocates and congressmen alike providing close analysis of the passages that maybe we should pay attention to what if over
00:06:43
texts like this poem it's hard to read by Edna st. Vincent Millay teachers can involve invite students in to annotate not with words but with pictures the passages that inspired them
00:06:56
these annotations should be able to exist in layers different layers for different purposes groups if you will private public or read-only for special
00:07:06
member communities would have teams of hundreds of climate scientists for instance could work together in these in their own group layers to critique the factual coverage of a reporting on
00:07:20
global warming it turns out they are this article has been annotated together in a project called climate feedback that has received quite a bit of attention but what if every community that is deeply informed about a
00:07:32
particular subject could run their own distinct group layers over the web no it's not all about Noble or academic pursuits here's a recipe for turkey soup that I like I've made a note on it with
00:07:46
some changes to the ingredients that I've arrived at at the bottom I pasted a YouTube video on Gordon Ramsay's most efficient way to dice an onion what if any user could search their annotations
00:07:59
across the web or the public annotations of any other user or group and against facets like tags and URLs with wildcards all these examples are not only possible
00:08:10
but being done now at scale hypothesis the nonprofit that I work work at hosts an endpoint of this open source software that services hundreds of thousands of users who've made about four and a half million annotations
00:08:24
growing rapidly this user profile pages for Heather on our team who has now made tens of thousands of annotations the vast majority of which are her own personal notes you can go to our website and
00:08:37
download a Chrome extension so that you can join the conversation free no ads no selling of user data ever it's all open source anybody can download it from
00:08:48
github for kit modify it fix it run an end point for them for their own purposes pull requests are welcome but we can do more what if every one of the 19 million books and texts in the
00:09:01
internet archive could be the launch pad for a book club about its contents like this copy of HG Wells World brain or this original scan of Doug's 1962 paper well what about every book available
00:09:15
through Amazon on the Kindle yeah it may take a lot of convincing to get Amazon to build this in natively to the Kindle but if you use the web reader which gets into you which is in your space in the browser you should be able to lay
00:09:27
annotations over those books there are still a few tricky bits to get this particular use case to work but we've got an early prototype that we're experimenting on what if annotation clients were able to not just connect to
00:09:41
a single monolithic server but to multiple services spin one up point people to it and let them to subscribe to it if they're interested there are million subreddits with billions of conversations but from any given page on
00:09:54
the web you'd never know that creating an open scaffold for this to work is one of our biggest goals this year to give you a more immediate example this week when I was reading up on my history to
00:10:07
prepare for this presentation I was finding the old original famous articles that represent the history of digital annotation the way that we mean it as a means of collective note-taking / knowledge I kept seeing a reference to a
00:10:19
1965 article of Ted's so we went to find a free copy but I couldn't find so I paid ACM the fifteen dollars for a fresh PDF and I downloaded it I began reading the article which it turns out is also
00:10:32
the one where Ted first coins the term hypertext and which if you haven't read it yet I highly recommend no not the first one I'll change my annotation if
00:10:46
you haven't read it yet I totally recommend it and for no surprise for anybody who's familiar with Ted's writing it's beautifully written to boot but I quickly wanted to be able to take notes and tag various passages so that I
00:10:59
could come back to them later so I dragged the PDF from my desktop to my browser and turned on the annotation clan lo and behold I discovered that it was already annotated but not because those annotations shipped with a copy I
00:11:11
downloaded from ACM you see every PDF comes with a unique hash which is a sufficient identifier for the annotation engine to search against the server returns the annotations for that fingerprint and the client reenters the
00:11:24
in real-time one of the annotations made last year by Roland Legrand is a note to read about project Xanadu and to watch the Werner Herzog documentary lo and behold that includes a segment on Ted I
00:11:39
added my own annotation anchored to this term zippered lift which I'm Ted Lemieux know if I get this wrong which was the first mention of a term later known as as analogical storage as analogical
00:11:52
storage and then transclusion a concept which needs the ability to point specifically into our any arbitrary place on the web like an annotation in this article Ted also reflects on Van
00:12:05
avars musing musing in as we may think which is probably the first real mention of digital annotation the way that we mean it I had an annotation to his quote and linked it back to an annotation at Van
00:12:18
avars original article at the Atlantic magazine website which over the last four years or so has accumulated about a hundred and fifty-seven annotations from dozens of people including this one by Jeremy over Van avars original mention
00:12:32
of marginal notes which simply says boom I added one replied to Jeremy there and added an annotation that points back to Ted citation a bi-directional length
00:12:46
what would Doug's andum annotations over Vannevar x' articles have been and how could those as Vint noticed have pointed back to us as readers from the future
00:12:59
I know Gardner has created a Twitter account for his project called Engelbart but maybe there ought to be an annotation account called Engelbart which can annotate the key things on
00:13:11
documents at the time that might have inspired Doug's work from these early mentions the notion of digital annotation evolved in Tim's 1989 proposal vague but exciting that we now
00:13:25
know as the web he suggested that for CERN's purposes you would need to be able to annotate links as well as nodes or pages for years later in 1993 mark
00:13:36
Andresen and Eric bina built annotation right into the browser and called for some guinea pigs to test it it had a very simple feature set one group for annotation server one annotation server per mosaic session no security except
00:13:50
through obscurity and to quote it does work in mosaic 1.1 has built-in client-side support they left it on for a couple months and then after considerable debate turned it off so they could focus back on the browser
00:14:04
that was the last time a browser has had native clients of side support for annotation over the 25 years since then there have been over 70 projects by our count which have tried to bring some
00:14:17
flavor this over the web some of them you may have heard of third voice reframing genius but none of these have caught on I asked why and spent about a year interviewing people back in 2010
00:14:30
before hypothesis why from their perspective from the founders perspective we they hadn't quite crystallized and the simple answer came that's in order for this to be effective
00:14:42
it needs to be open based on standards and not captured by any one single team or implementation in 2014 a number of us got together and started working group at the w3c with Tim
00:14:55
support last February that recommendation for web annotation which includes the formal data model the protocol and the vocabulary was ratified by w3c membership this allows
00:15:06
interoperable clients and services to be billable very much like the web itself or email or IRC since then we put together a coalition of groups which
00:15:20
have pledged to bring annotation natively to their platforms which include a lot of open public publishers folks like JSTOR and plus and and Wiley and folks you may have heard of and
00:15:33
we're looking forward to the future so our mission to enable a conversation over all knowledge thank you for your time that's great it before we get just started it turns
00:15:51
out we actually know what Doug's annotations on the article would be the Computer History Museum has Doug's copy out of the original Atlantic and there are pencil notes in the margin so
00:16:04
somebody in the Computer History Museum Jake could actually follow up and annotate the online version of the article with
00:16:13
Doug's comments that would be cool okay please go ahead in a spirit of kind of
00:16:30
rich complex problems and social interaction I woke up this morning having no idea what I was going to talk about today so by way of introduction
00:16:49
it's really impossible it takes me ten minutes to explain this to people who know me well so in addition to Doug who inspired me about collective
00:17:01
intelligence I also was greatly influenced by Horst Rittel a professor at Berkeley who had come up with a
00:17:14
system to help structure rich complex conversations called Ibis and so the
00:17:27
other and riddle also came up with the term wicked problems just I curiosity how many people have heard of a wicked
00:17:37
problem okay so I'm gonna talk about that a little bit more it doesn't have anything to do with morality it has to
00:17:52
do with the ability of the people working on the problem to make sense of what they think the problem is so
00:18:05
it's pretty meta so I was just blown away by this idea of Ibis as a structure of questions and ideas and arguments as
00:18:19
a way to kind of a kind of a rhetorical grammar a conversational structure for dissecting and displaying really rich
00:18:34
hard conversations among any number of stakeholders at any level of difference and having it make more sense having the
00:18:47
display of this conversation in Ibis at least if nothing else lower the tendency to have misunderstanding turn into
00:19:00
hostility so that's what I did for a couple decades and and now I'm a
00:19:15
handyman in Napa so yeah now you know okay so I'm sure you've heard the old joke about the drunk guy who's
00:19:27
staggering around under a streetlight he's looking down after a little while another guy comes along and asks him what's he looking for my case says the
00:19:43
drunk the other guy looks down and starts looking around is this where you lost them he asked casually no I lost him over there he
00:19:56
says pointing off into the darkness but the lights better over here so I used to think this was a funny joke this is cute I now think it's possibly
00:20:12
one of these excuse me one of the most important parables in modern civilization because the keys right the
00:20:30
keys represent something that is potentially of immense value to humanity and to the planet when you're talking at
00:20:42
Doug's level of commitment and the drunk guy I don't think you're going to like this but the drunk guy represents the
00:20:57
computer industry looking mainly under the streetlight of profit and funding creating technologies that are conceptually an easy reach within the current state of the practice
00:21:12
and so the point is we we mostly build tools that do what computers can do easily not necessarily what humans need
00:21:30
and this I think is the the summary and also the the diagnosis for Doug's Doug's
00:21:41
work and his frustration Allen Kay's vision of interface as a learning environment what a great example of
00:21:58
something that is of course and yet who tries to do that and who's been successful at finding what the user is
00:22:14
confused about because the language or the arrangement of things or whatever is not what they're expecting and there's no model of this user so it the system just sits there and stares at them and
00:22:29
there's no possibility of learning how to use this thing that the damn thing you let unless you ask someone or spend a couple of hours looking it up
00:22:38
someplace so okay back to the metaphor occasionally someone comes along who's so devoted to finding the real key that
00:22:52
they're willing to look in the darkness Doug was such a person you know a visionary someone with good night vision but it can be lonely out
00:23:12
there looking for the keys stumbling around in the darkness so the smart visionary moves in closer to the
00:23:25
streetlight where the lights just good enough that you can do something and maybe some of the other sort of nocturnal souls that are looking for
00:23:40
these keys can find you and you all can go to this kind of space at the edge of what's easy and obvious and work on it
00:23:57
sorry no flashlights allowed actually flashlights are I'll come to that later it fits in the metaphor
00:24:09
so Doug and his band of visionaries sensed that these keys there they're not
00:24:24
it's not a just a quick let's look around out here in the darkness thing to find the keys they're there they might be off in the in the deep underbrush
00:24:35
over here so they decided to do is a very clever thing they built a model of what the keys might look like and that
00:24:51
was the demo 68 so the point of the demo in this perspective is to draw out the other searchers into the darkness where you
00:25:12
don't know what's going on and you don't know what the problem isn't and by the way there's no funding where the real
00:25:21
Keys had to be so I was one of those searchers I wasn't on Doug's team but I had already found another visionary this
00:25:34
guy Horst riddle that I mentioned and he had created the Ibis system and coined the term wicked problems and those two things just really sung to me and I wanted to build a bridge between what
00:25:48
riddle was doing and what angle bar was doing so I set off down that path and [Music]
00:25:59
like I say now I'm a handyman so it's the challenge as I see it the challenge is to somehow equip ourselves embolden
00:26:15
ourselves to go into that darkness together where the problems aren't simple and in fact and let me just give you a little bit of a sense of what
00:26:28
makes a wicked problem because not everybody raised their hands I don't know if you notice on a wicked problem
00:26:38
what the problem is is a is a debate there's no agreement about it in fact it can be a place of real conflict it's a new problem
00:26:52
no one's ever sold anything like it so there are no experts there's nobody you can't get on the phone and hire someone from a consultancy to come in and solve the problem for you generally there's a
00:27:04
long history or at least some history of litigation around this problem it's one of the key markers no litigation probably not wicked the biggest stakeholders are not
00:27:20
at the table how come because they're betting that whatever this beautys group of people come up with as a solution to the problem fails because the biggest
00:27:33
stakeholders are committed to the status quo and it's not in their interest to have some bunch of yo-yos come in and come up with a solution that gets any momentum
00:27:50
sorry I can see I can just feel the gravity settling in and I know this is I'm not waving the technology flag up
00:28:04
here if you've noticed so I'm gonna sort of stop there okay and oh and Jeff I'm gonna leave it to you to kind of bring people back maybe
00:28:20
I'll try I'll try okay great cheer us up and lead us on okay okay great so so let me stay on that wicked problem for a minute
00:28:38
changing my thoughts around as I'm doing this Pat Lincoln at SR I had a had a great way of talking about one of those ideas so you know Doug often talked
00:28:53
about working on the world's great problems and and I'm sure if I ask in the audience what are the big problems are facing us I'd get answers it's
00:29:05
climate change they're a big problem is is disease and these are the world's big problems Doug's big wicked problem was how do we improve the way we solve wicked problems
00:29:19
and I think it has a lot of these characteristics and in fact I went out I tried with Doug during the 90s to get funding from companies and stuff like
00:29:32
this and they were all enthusiastic about working on their problem they did not want to work on how to improve the way they work on their problem it's it
00:29:45
was quite an interesting problem and it still is so he'd appreciate the wicked problem issue a lot I had an idea when Dan was talking about I wanted to go
00:29:57
into bootstrapping about that but there was another idea that hasn't been talked about here at all that Doug had Doug saw text and the manipulation of text is the
00:30:08
thing we could do in 1968 his insight actually going back into the late 50s the fundamental insight was was all of the media that we deal with in the world
00:30:22
is going to become digitized and audio even smell although we used to argue about that a lot but and as it as our as video and audio becomes digitized
00:30:36
there's no need to treat it the way that we treat it you don't have to manipulate text like you were manipulating cut out things to put a paper together and you don't have to compose music that can be played by people and you don't
00:30:49
have to have videos that that have people in them and and as we go through that transition we're actually going through part of Doug's vision of going from from what we could do with
00:31:03
mechanical media into into a world that we can live in where everything is digitized Allen had picked up on this you don't Allen is an expert organist
00:31:17
and when you can when you can score your music using computers and you can play it on computers you don't have to stick this stick to a four-part fuse you can have seven part fuse you know and do
00:31:31
stuff that you couldn't do I want it let's we only have three minutes with with hypothesis I want to dive in three
00:31:42
minutes into the ABC example does the hypothesis team use hypothesis in any way when they're building hypothesis we we use it internally to take notes and
00:31:58
for many different kinds of things we don't use it necessarily as a fundamental part of the software development process right right but we use it as part of maybe the conceptual thinking okay how we do it so there we
00:32:12
have a good example of the first stage of bootstrapping the this sea level thinking that Doug would talk about the answer may be negative here is is have
00:32:25
you ever gone through a process if using hypothesis to improve your sort of your long-range plans for hypothesis or their
00:32:36
way to improve the way you're improving hypothesis discussions on that level I mean to the extent that when the thinking itself around product you know
00:32:51
what our plans our forecasts and things like that are appropriate for the kind of annotation that we're doing yeah although you know we use a lot of Google Docs to Google Docs has a great little
00:33:02
annotation capability in it and I really I mean the thing that struck me this morning and listening to the conversation you guys were having about the making of of the demo was what an
00:33:17
extraordinary experience it was to be building all the tools necessary from the the meta compiler that needed to be written first in order to to be able to to get there and the kinds of
00:33:29
opportunities that projects that are susceptible Ornette need that kind of ground-up architecture are pretty interesting these days we have so many open-source modules that were grabbing
00:33:41
you know that happened to be on github that we just kind of bolt together along with some other code to do the things we're doing it so it's a different in some ways a different I'm sort of pushing on this really to demonstrate
00:33:52
that these ideas that Doug had are it's easy to get through the a level work and the B level work but getting into the really deep sea level work and the bootstrapping process that goes on behind it is even extremely difficult
00:34:06
for people that are building the right kind of tools and know what they're doing okay I think we're done and we're on time again [Applause]
End of transcript