Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • May 27 06:08
    User @Mitzi-Laszlo unbanned @in1t3r
  • May 23 06:49
    @Mitzi-Laszlo banned @in1t3r
  • May 16 09:49
    @Mitzi-Laszlo banned @mediaprophet
  • Feb 01 22:04
    User @melvincarvalho unbanned @namedgraph_twitter
  • Feb 01 21:49
    @melvincarvalho banned @namedgraph_twitter
Tim Berners-Lee
@timbl
"Git is effective because it requires rigour to use, which serves as a filter. But in this case it is being used, like the other forms, to produce transitory prose rather than articulated and finely reusable information.” — indeed
More apprpriate is perhaps soething like dokieli where one person at a time edits, and everyone can annotate
Tim Berners-Lee
@timbl
I feel we have a lot of tools with solid which we need to connect together to form solutions. Also we need to work on each one to get ot more polished.
David H. Mason
@vid

Dokieli is fantastic, but from what I know prose-document focused. A few projects are implementing linked data components, which allow choosing other pods or information sources, and that would be a basis of distributed project development. Original human text is important, of course, but can be a problem when it's overwhelming. There should be every opportunity and incentive to include dynamic well defined views. I'm doing this with components embedded in markdown (mdx), and smartdown has an implementation based on this approach as well. The embedded approach can help solve the polish issue, since people can embed components and focus on their own style.

One immediate outcome is the regular This Week in Solid bulletins could be largely automated, with curated queries from different sources. Embedding task tags across repos would allow a very fine overview of ongoing activities.

Mark Hughes (happybeing)
@theWebalyst
@Mitzi-Laszlo thanks. I was taking about the joining and finding out process - because it says to join a panel you submit a PR. Maybe you can join a channel before that? If so, perhaps add that to the info so the first step doesn't seem like coming a repo and creating a PR. To me that is unclear and onerous to discover compared to requesting an invitation to join a discussion group. Anyway, I've raised the issue.
Sarven Capadisli
@csarven
@vid Seen https://dokie.li/media/video/dokieli-sparqlines.webm ? That uses a SPARQL endpoint with statistical linked data to build an inline chart in SVG. Some write-up at https://csarven.ca/sparqlines-sparql-to-sparkline
Justin Bingham
@justinwb
that is slick @csarven :+1: :+1:
Mike Adams
@mikeadams1
@csarven that is slick indeed! I am just getting into sparql, and linkeddata, and that Query engine is fantastic.
Mark Hughes (happybeing)
@theWebalyst
Amazing @csarven
David H. Mason
@vid
image.png

@csarven @csarven the context is "dogfooding," using our own tools to prove they work and practically develop them. In this case, to support the most people, start with the top level, the main project summary; This Week in Solid. Write simple importers for everything related to Solid news and development to the degree if it's not a query response it doesn't exist. That would require all sorts of different queries in the early days, from github organizations and repos, forums, chat, etc. But all those APIs exist.

Consolidate them to one linked data store, using a Curate component that allows an editor to filter them (of course saving the result in linked data, so other people can develop with that list). Not only does it become progressively easier for any person to get a live top-level view, useful APIs (github, etc) are getting linked-data-ified, with highly reusable curated data as another product. Aaaaand it's an excellent demonstration of decentralized social linked data that solves real-world problems.

Sarven Capadisli
@csarven
@vid You don't have to convince me about dogfooding :)
That can all be retrieved, stored, queried. Would you like to do it?
Sarven Capadisli
@csarven
eg. https://linkedresearch.org/ldn/tests/summary lists LD datasets referring to individual implementation reports which are themselves in LD. The whole mechanism to create and then later fetch, build summary is driven with LDN.
So, sign me up as to where I can deposit my dogfood
Henry Story
@bblfish
It is very much worth looking to Category Theory as an ally of the Web. See a number of my recent Tweets, and this article from Forbes as an introduction The future Will be Formulated using Category Theory
Henry Story
@bblfish
That article by Forbes is a little too hyped for my taste. But that is rare. Most articles in that space are too abstract. But there is a growing number of contributions that are very practical such as Category Theory for Programmers or for those of you from MIT Category Theory for Scientists that is really about databases (and there are some articles to come out tying graph databases to RDF via CT).
David H. Mason
@vid
I would like to do it, but I have a full-time pursuit, that I would like to bring into this world, as it becomes ready. So I am waiting for smarter people like yourselves to build the pieces. (;
I have worked with related projects that prove the value, though I know it's just scratching the surface. But just for this one piece, what I mean by dogfooding is specifically making the Solid weekly reports the product of generated queries, with some human curation, since that should drag everything else along into a cohesive system. For people for myself who can't live and breathe this work it would make it much easier to understand and follow development. But that would have to be a dedicated project on its own, and I couldn't lead it. I wish I could!
Sarven Capadisli
@csarven
@vid I generally agree with you on "nice to haves" however I'm not sensing a particular end goal or a limitation in for example with the weekly reports. FWIW, I find the weekly reports useful as is - it is well organised for human-consumption. It is a summary. It has links that can be followed. Certainly we can do all sorts of things around that or even autogenerate stuff, but I doubt that we are going to draw line in the sand right now and say "if your Solid activities are not 'queryable' it doesn't exist". Of course things exist on the Web by merely having a URI. Query is not some dependency or ought to be. If you'd like some particular features, then why not propose what they are. Or possibly propose which tools you are willing to use in order to take better advantage of the weekly reports. We can still move forward while using Gitter, GitHub, and a host of other things as we learn more, shape and improve the ecosystem.
Having said that, if people want to send a simple notification about their activities to a dedicated Solid inbox eg. https://solidproject.org/inbox/this-week-in-solid/2019-08-30/ or whatever... indeed we can list all of the activities automagically. That's about as simple as it gets right now I think.
<> a as:Announce ;
  as:actor <the agent that's sending the notification> ;
  as:object <the thing that's of interest to the Solid community> ;
 ... and so on...
Sarven Capadisli
@csarven
Then we can do something as basic as https://dokie.li/?open=https://rhiaro.co.uk/2018/12/ listing activities or notifications...
David H. Mason
@vid
I don't mean that it would be immediately the case that something "doesn't exist," though it may be factually. It's just to reduce the burden to collect information using, well, decentralized social linked data mechanisms, so that editors can focus on the aspects that specifically require human apparatus past copying and pasting. And it would better serve understanding specific macro and micro details without having to follow many links in arbitrary formats. But it's just a suggestion to the question of how to make everything easier to track. I'll try to work on it as I can, of course.
Scott McWhirter
@konobi
@vid reducing overhead and cognitive load for summarization is definitely a worthy goal
Scott McWhirter
@konobi
@bblfish have you happened to see the N J Wildberger videos on youtube at all?
Dorian Taylor
@doriantaylor
TIL you can access gitter with irssi
now i might actually participate
Scott McWhirter
@konobi
There's an unexpected face... waves
Dorian Taylor
@doriantaylor
how's it going scott
i've been lurking for most of the year
i just find the gitter web interface to be too much clutter
Scott McWhirter
@konobi
So-so... trying to gather thoughts together while dealing with neuropathy is a real PITA.
Dorian Taylor
@doriantaylor
yikes
you still in vancouver?
Scott McWhirter
@konobi
yeah
Henry Story
@bblfish
@konobi I have not seen N J Wildberger's videos on youtube. Any reason I should have? (There's a lot to learn out there)
Scott McWhirter
@konobi
@bblfish he covers linear algebra in a pretty unique way (among other topics)... which has some relevance to sparql
(geometric/intuitive thinking)
Henry Story
@bblfish
Could be good. I thought I'd never came to use my A level (S-level actually) algebra again, and then with category theory I discovered that programming is algebra. So I should do a refersher course.
Dorian Taylor
@doriantaylor
i've been plodding through physics and math lectures myself for the last few years, it's how i knock myself out at night
Scott McWhirter
@konobi
@doriantaylor sounds familiar ^_^
Dorian Taylor
@doriantaylor
most recently gilbert strang's ocw
now i'm onto 3b1b but there are too many ads
(when looking at it on a phone at least)
i should compile a list
Scott McWhirter
@konobi
3b1b++
Henry Story
@bblfish
If you want to get a quick overview of how CT ties into programming, check out the talk by Bartosz Milewski The Maths Behind Types.
Dorian Taylor
@doriantaylor
yeah i watched all the bartosz lectures earlier this year
i keep meaning to order a copy of the book