Re: How the father of the World Wide Web plans to reclaim it from Facebook and Google

On 17 August 2016 at 10:40, Martynas Jusevičius <martynas@graphity.org>
wrote:

> My 2 cents: Linked Data is great, but what it lacks is a theoretical
> model below it, on which other efforts can build. Of the W3C specs,
> only RDF has semantics, while SPARQL has an algebra. Why not LDP? That
> has produced a myriad of LD stacks with duplicating features and low
> interoperability, both in terms of composition of software libraries
> and run-time LD calls, mostly on the write side.
>
> Someone has said earlier on this list, that (software) engineering is
> not science. Well, maybe we should turn it into science then. And use
> it to produce one generic Web API (read-write Linked Data), instead of
> the "API economy" which we currently have.
>
> Basically, Linked Data should have provable semantics. We think we
> have found a declarative way to do it, which actually brings Linked
> Data closer to the original ontology-driven Semantic Web vision. But
> it involves SPARQL, which many Linked Data people seem to have an
> aversion for (yet many of the same people champion JSON, which is an
> immaterial and orthogonal implementation detail in this big picture).
>
> We call the approach Linked Data Templates, and are currently working
> on its semantics. Please take a look:
> https://github.com/AtomGraph/Linked-Data-Templates/blob/
> master/XML%20London%202016%20paper/Linked%20Data%20Templates.pdf


I'm all for the declarative approach!

I think what some people have alluded to here is that to give technology
the best chance, you have to find a sweet spot.

It has to be generic enough to solve a large number of use cases and
practical enough to get a developer and user base.  It needs a development
team behind it and some specification work / documentation.  The main issue
I see is that there are very few resources in this area.  One reason it's
taken more than 10 years (or 20?) to get as far as we have.

Inevitably it's about finding compromises to get to a spec good enough to
solve the important use cases (with large coverage), and with enough mind
share and developer momentum to be worth the time investment.  Competing
stacks are something we want to minimize simply because we lack developers
to really make a good go of each.

LDP I think was good in that we needed to webize the file system, and it
was a pretty good attempt at that.  Of course, many people will have
differing views on what LDP is for.  This is really useful and Solid simply
adds access control to that, because a file system without permissions is
less useful.

But what we have in linked data now is just about a stack capable of a
whole new generation of apps.  Yes we are held back by legacy technology
which means developers have to work that much harder to make the case.  But
as soon as people start to see things working, and what can be done, it is
possible to solve our one major problem which is lack of developers.  Id
encourage those interested in LD to try and collaborate more and work on
interop.  Of course this is really really hard, because in a distributed
environment there are more moving parts, and this increases the testing
complexity not double but an order of magnitude.  And pretty much no one
has time to do that kind of testing -- we need a solution here -- and for
years I've not been able to think of any, other than to suggest having a
'head of integration' in teams.

Until then I think we really need to try and get at least one project to
the degree of maturity where it can challenge and overtake the best we've
seen on web 1.0 + 2.0 and bring it to a next generation distributed
paradigm.  RDF is great.  Linked Data is great.  SPARQL is a great query
language.  LDP is a great tool to work with file systems, of file system
like things.  Solid is a great tool to add identity and permissions, and
realtime (which sadly isnt working on android right now).  We dont really
need much more than this, just encourage people to build apps!


>
>
> Best,
>
>
> Martynas
> atomgraph.com
>
>
> On Wed, Aug 17, 2016 at 4:24 AM, Timothy Holborn
> <timothy.holborn@gmail.com> wrote:
> > Melvin / Henry (or TimBL)
> >
> > Can someone ask (or respond) and tell me what the key principles are that
> > TimBL wants to achieve in producing SoLiD (like?) alternatives.
> >
> > What are these key qualities...
> >
> > I imagine he would / could summarise it in a page or so, as he has done
> with
> > other concepts in the past.
> >
> > I do not believe he has a rigid view that SoLiD is the only path for the
> web
> > into the future.  A comment i am reminded of, is one of creating pieces
> and
> > not forcing the everything to be used; but hoping counterparts can and
> will
> > be.
> >
> > I see work that's been done over a VERY long period of time; and i think
> the
> > semantic inferencer that has somehting like HTTP signatures protecting
> > algorithms shared using something like linked-media-fragments to services
> > were people are storing their private and sensitive media objects that
> they
> > want to be processed by algorithms produced by incredible scientists
> around
> > the world - well, that kinda stuff is amongst the 'to-do' list IMHO.
> >
> > TimH>
> >
> > On Wed, 17 Aug 2016 at 09:31 Melvin Carvalho <melvincarvalho@gmail.com>
> > wrote:
> >>
> >> On 16 August 2016 at 20:23, Kaliya IDwoman <kaliya-id@identitywoman.net
> >
> >> wrote:
> >>>
> >>>
> >>> On Tue, Aug 16, 2016 at 5:51 AM, Adrian Hope-Bailie
> >>> <adrian@hopebailie.com> wrote:
> >>>>
> >>>> What is the business case for a service provider to adopt Solid?
> >>>
> >>>
> >>> 1) first off I'm super skeptical of any project that is
> >>> university/research based it is notoriously difficult to get those to
> escape
> >>> the lab as it where. Everyone has incentive to "publish" for their
> >>> degrees/professorships - zero incentive to make a usable, market
> >>> worthy/ready product (I don't just mean in a business way relative to
> market
> >>> but adoptable in the marketplace of tools and software)
> >>>
> >>> 2) The Personal Data Ecosystem Consortium that I founded in 2010
> >>> http://www.pde.cc has a whole range of companies that have been
> working on
> >>> similar technology and ideas for well over 5 years. So it isn't new -
> the
> >>> ideas around personal data stores/banks etc and putting people at the
> center
> >>> of their own data lives go back at least to Johannes'  Ernst work (See
> the
> >>> top of my twitter for a diagram he drew in 2005-6.  And the Augmented
> Social
> >>> Network White Paper which itself and antecedents in other work.
> >>> http://asn.planetwork.net
> >>>
> >>> 3) Please show me what Tim has lead that has gotten to market besides
> >>> HTML back in the day?
> >>
> >>
> >> Skepticism is healthy.  But can sometimes be overdone.
> >>
> >> Tim didnt just get html to market.  He also created the first browser
> >> (editor).  He created HTTP.  He created.  He created the first web
> server.
> >> And after that he created linked data.  And now Solid.  This is all
> really
> >> one project known as the world wide web.
> >>
> >> Fun fact: when presenting these things to the hypertext conference when
> it
> >> was all working, the paper was rejected from the main conference and
> only
> >> allowed "poster track"
> >>
> >> See:
> >>
> >> https://www.w3.org/2004/Talks/w3c10-HowItAllStarted/?n=16
> >>
> >> Simple fact is that Tim thought about the web for 2 decades before
> >> releasing it.  Almost no one got it then.  Solid is the conclusion of
> that
> >> work, and almost no one gets it now.  My hope is that people will start
> to
> >> appreciate it when they see it in action! :)
> >>
> >>>
> >>>
> >>>
> >>>>
> >>>>
> >>>> Why would Google, Facebook or anyone that build's their business on
> user
> >>>> data choose to let users take that away?
> >>>
> >>>
> >>> They don't have a choice because the European regulatory framework the
> >>> General Data Protection Regulation that comes into force in 2018 is
> >>> mandating it.
> >>> You also have a whole group of companies working on building businesses
> >>> around this premise and one finally finally got funding -
> >>> https://techcrunch.com/2016/06/30/digi-me-bags-6-1m-to-
> put-users-in-the-driving-seat-for-sharing-personal-data/
> >>> Meeco https://meeco.me/ from Australia is doing awesome work (Both
> there
> >>> and in the UK) as is MyDex https://mydex.org/
> >>>
> >>>
> >>>>
> >>>>
> >>>> Who will offer users a comparable service to these silos that attracts
> >>>> them away but adopts Solid and can still make enough money to survive
> >>>> competing with the biggest tech companies in the world?
> >>>>
> >>>> The point is not whether or not the architecture is easy the point is
> >>>> whether it has the potential to make anybody any money because if it
> doesn't
> >>>> then I think you will have a hard time persuading people to use it, no
> >>>> matter how well it scales.
> >>>
> >>>
> >>> We have to really get into the weeds of figuring how value flows in
> these
> >>> networks to make it work for the parties involved and be sustainable
> in the
> >>> long run.  It will take way more then "architecture".
> >>>
> >>>
> >>> If you all want to dive into some of the nitty gritty I invite you to
> the
> >>> Internet Identity Workshop - http://www.internetidentityworkshop.org
> >>>
> >>> :) Kaliya
> >>>
> >>>
> >>>>
> >>>>
> >>>> On 15 August 2016 at 14:11, Melvin Carvalho <melvincarvalho@gmail.com
> >
> >>>> wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>> On 15 August 2016 at 14:08, Timothy Holborn <
> timothy.holborn@gmail.com>
> >>>>> wrote:
> >>>>>>
> >>>>>> Solid isn't finished yet.
> >>>>>
> >>>>>
> >>>>> Solid is at version 0.6 rather than 1.0.
> >>>>>
> >>>>> But I dont really know what more can be added to it to get it to
> v1.0.
> >>>>> Im using it on a daily basis and it works fine.  Some people are
> >>>>> perfectionists I suppose :)
> >>>>>
> >>>>> In any case its IMHO light years ahead of where the rest of the web
> is,
> >>>>> even if you only take small parts of it and use it.
> >>>>>
> >>>>> You can also argue that solid will never be finished, in the sense
> >>>>> that, the web will never be "finished".
> >>>>>
> >>>>> Its definitely something that can be used today.
> >>>>>
> >>>>>>
> >>>>>>
> >>>>>> On Mon, 15 Aug 2016, 10:07 PM Melvin Carvalho
> >>>>>> <melvincarvalho@gmail.com> wrote:
> >>>>>>>
> >>>>>>> On 15 August 2016 at 11:50, Adrian Hope-Bailie
> >>>>>>> <adrian@hopebailie.com> wrote:
> >>>>>>>>
> >>>>>>>> From the article: "The question is whether architecture will be
> >>>>>>>> enough."
> >>>>>>>>
> >>>>>>>> The answer is no.
> >>>>>>>> We live in world where few ideas succeed without a strong business
> >>>>>>>> case. The architecture is the easy part.
> >>>>>>>
> >>>>>>>
> >>>>>>> Architecture is deceptively difficult to get right.  The vast
> >>>>>>> majority if systems start to fall over as they scale.  The web and
> REST are
> >>>>>>> two architectures that buck that trend and just get stronger as
> they scale.
> >>>>>>>
> >>>>>>> Solid is the next evolution in that architectural trend, imho,
> >>>>>>> because it simply embraces the points that made the web great, and
> extends
> >>>>>>> it a little bit, while being 100% backwards compatible.  Right
> now, it's the
> >>>>>>> only system that I know of, with this property, in fact, nothing
> else is
> >>>>>>> close.  So this in itself, the ability to scale to billions of
> users, is a
> >>>>>>> business case.  Quietly facebook adopted the social graph approach
> to the
> >>>>>>> web, and web architectural principles with their graph protocol,
> and also an
> >>>>>>> implementation of WebID.
> >>>>>>>
> >>>>>>> I think what's true is that few ideas succeed, because simply, we
> >>>>>>> have a lot of ideas and a lot of competition.  Having a business
> can help,
> >>>>>>> but the right architecture is the magic sauce to get through those
> >>>>>>> scalability barriers.
> >>>>>>>
> >>>>>>> I personally think Solid is the business opportunity of a lifetime,
> >>>>>>> perhaps even bigger than the first web.  Im certainly investing on
> that
> >>>>>>> basis.
> >>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On 14 August 2016 at 10:49, Timothy Holborn
> >>>>>>>> <timothy.holborn@gmail.com> wrote:
> >>>>>>>>>
> >>>>>>>>> Hi Anders,
> >>>>>>>>>
> >>>>>>>>> I'm using this email to respond to both [1] in creds; in addition
> >>>>>>>>> to the below, with some lateral considerations.
> >>>>>>>>>
> >>>>>>>>> See this video where Mr Gates and Mr Musk are discussing in China
> >>>>>>>>> AI [2].
> >>>>>>>>>
> >>>>>>>>> I haven't fully considered the implications, whilst i've
> certainly
> >>>>>>>>> been considering the issue; i have not fully considered it, and
> as modern
> >>>>>>>>> systems become subject to government contracts as may be the
> case with
> >>>>>>>>> enterprise solutions such as those vended by IBM [3], may
> significantly
> >>>>>>>>> lower the cost for government / enterprise, in seeking to
> achieve very
> >>>>>>>>> advanced outcomes - yet i'm unsure the full awareness of how
> these systems
> >>>>>>>>> work, what potential exists for unintended outcomes when work by
> >>>>>>>>> web-scientists[4][5] becomes repurposed without their explicit
> and full
> >>>>>>>>> consideration of the original designers for any extended use of
> their works,
> >>>>>>>>> what the underlying considerations are by those who are
> concerned [6][7] and
> >>>>>>>>> how these systems may interact with more advanced HID as i've
> kinda tried to
> >>>>>>>>> describe recently to an audience here [8] and has been further
> discussed
> >>>>>>>>> otherwise [9] [10].
> >>>>>>>>>
> >>>>>>>>> I'm a little concerned about the under-resourcing that seems to
> >>>>>>>>> plague Manu's / Dave's original vision (that included WebDHT) to
> the
> >>>>>>>>> consultative approach that i believed had alot of merit in how
> it may
> >>>>>>>>> interact with the works of RWW at the time (alongside WebID)
> which have al
> >>>>>>>>> progressed, yet, not seemingly to a solution that i think is
> 'fit for
> >>>>>>>>> purpose' in attending to the issues before us.
> >>>>>>>>>
> >>>>>>>>> I have considered the need for people to own their own biometric
> >>>>>>>>> signatures.  I have considered the work by 'mico-project'[11]
> seems to be a
> >>>>>>>>> good supporter of these future works, particularly given the
> manner in which
> >>>>>>>>> these works support LDP and other related technologies...
> >>>>>>>>>
> >>>>>>>>> But the future is still unknown, and what worries me most; is
> those
> >>>>>>>>> who know most about A.I. may not be able to speak about it as a
> citizen or
> >>>>>>>>> stakeholder in the manner defined by way of a magna carta, such
> as is the
> >>>>>>>>> document that hangs on my wall when making such considerations
> more broadly
> >>>>>>>>> in relation to my contributory work/s.
> >>>>>>>>>
> >>>>>>>>> i understand this herein; contains an array of fragments; yet, am
> >>>>>>>>> trying to format schema that leads others to the spot in which
> i'm
> >>>>>>>>> processing broader ideas around what, where and how; progress
> may be
> >>>>>>>>> accelerated and indeed adopted by those capable of pushing it
> forward.
> >>>>>>>>>
> >>>>>>>>> I remember the github.com/Linkeddata team (in RWW years) wrote a
> >>>>>>>>> bunch of things in GO, which is what the IPFS examples showcase,
> and without
> >>>>>>>>> providing exhaustive links, i know Vint has been working in the
> field of
> >>>>>>>>> inter-planetary systems [13], therein also understanding
> previous issues
> >>>>>>>>> relating to JSON-LD support (as noted in [1] or [14] ), which
> in-turn may
> >>>>>>>>> also relate to other statements made overtime about my view that
> some of the
> >>>>>>>>> works incubated by credentials; but not subject to IG or
> potential WG
> >>>>>>>>> support at present - may be better off being developed within
> the WebID
> >>>>>>>>> community as an additional constituent of work that may work
> interoperable
> >>>>>>>>> with WebID-TLS related systems.
> >>>>>>>>>
> >>>>>>>>> Too many Ideas!!!
> >>>>>>>>>
> >>>>>>>>> (perhaps some have merit...)
> >>>>>>>>>
> >>>>>>>>> Tim.H.
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> [1]
> >>>>>>>>> https://lists.w3.org/Archives/Public/public-credentials/
> 2016Aug/0045.html
> >>>>>>>>> [2] https://youtu.be/TRpjhIhpuiU?t=16m26s
> >>>>>>>>> [3] http://blog.softlayer.com/tag/watson
> >>>>>>>>> [4] http://webscience.org/
> >>>>>>>>> [5] https://twitter.com/WebCivics/status/492707794760392704
> >>>>>>>>> [6] https://www.youtube.com/watch?v=tV8EOQNYC-8
> >>>>>>>>> [7]
> >>>>>>>>> https://en.wikipedia.org/wiki/Open_Letter_on_Artificial_
> Intelligence
> >>>>>>>>> [8] (perhaps not the best reference, but has a bunch of ideas in
> >>>>>>>>> it:
> >>>>>>>>> https://docs.google.com/presentation/d/1RzczQPfygLuowu-
> WPvaYyKQB0PsSF2COKldj1mjktTs/edit?usp=sharing
> >>>>>>>>> [9] https://www.youtube.com/watch?v=iTqF3w2yrZI
> >>>>>>>>> [10] https://www.youtube.com/watch?v=_x_VpAjim6g
> >>>>>>>>> [11] http://www.mico-project.eu/technology/
> >>>>>>>>> [12] https://www.youtube.com/watch?v=8CMxDNuuAiQ
> >>>>>>>>> [13]
> >>>>>>>>> http://www.wired.com/2013/05/vint-cerf-interplanetary-internet/
> >>>>>>>>> [14] https://github.com/ipfs/ipfs/issues/36
> >>>>>>>>>
> >>>>>>>>> On Fri, 12 Aug 2016 at 14:47 Anders Rundgren
> >>>>>>>>> <anders.rundgren.net@gmail.com> wrote:
> >>>>>>>>>>
> >>>>>>>>>> On 2016-08-11 15:16, Melvin Carvalho wrote:
> >>>>>>>>>> > Really good article, mentions Solid and other technologies.
> >>>>>>>>>> > WebID is mentioned by the author in the comments too ...
> >>>>>>>>>> >
> >>>>>>>>>> > http://www.digitaltrends.com/web/ways-to-decentralize-the-
> web/
> >>>>>>>>>>
> >>>>>>>>>> One of the problems with the Web is that there is no easy way
> >>>>>>>>>> letting a provider know where you come from (=where your Web
> resources are).
> >>>>>>>>>> This is one reason why OpenID rather created more
> centralization.  The same
> >>>>>>>>>> problem is in payments where the credit-card number is used to
> find your
> >>>>>>>>>> bank through complex centralized registers.
> >>>>>>>>>>
> >>>>>>>>>> Both of these use-cases can be addressed by having URLs + other
> >>>>>>>>>> related data such as keys in something like a digital wallet
> which you carry
> >>>>>>>>>> around.
> >>>>>>>>>>
> >>>>>>>>>> There is a snag though: Since each use-case needs special logic,
> >>>>>>>>>> keys, attributes etc. it seems hard (probably impossible),
> coming up with a
> >>>>>>>>>> generic Web-browser solution making such schemes rely on
> extending the
> >>>>>>>>>> Web-browser through native-mode platform-specific code.
> >>>>>>>>>>
> >>>>>>>>>> Although W3C officials do not even acknowledge the mere
> >>>>>>>>>> existence(!) of such work, the progress on native extensions
> schemes has
> >>>>>>>>>> actually been pretty good:
> >>>>>>>>>>
> >>>>>>>>>> https://lists.w3.org/Archives/Public/public-webappsec/
> 2016Aug/0005.html
> >>>>>>>>>>
> >>>>>>>>>> This is approach to decentralization is BTW not (anymore) a
> >>>>>>>>>> research project, it is fully testable in close to
> production-like settings
> >>>>>>>>>> today:
> >>>>>>>>>> https://test.webpki.org/webpay-merchant
> >>>>>>>>>>
> >>>>>>>>>> The native extensions also support a
> >>>>>>>>>> _decentralized_development_model_for_Web_technology_,
> something which is
> >>>>>>>>>> clearly missing in world where a single browser vendor has 80%
> of the mobile
> >>>>>>>>>> browser market!
> >>>>>>>>>>
> >>>>>>>>>> Anders
> >>>>>>>>>>
> >>>>>>>>
> >>>>>
> >>>>
> >>>
> >
>