- From: Melvin Carvalho <melvincarvalho@gmail.com>
- Date: Fri, 2 Sep 2016 16:14:03 +0200
- To: Kingsley Idehen <kidehen@openlinksw.com>
- Cc: business-of-linked-data-bold <business-of-linked-data-bold@googlegroups.com>, public-rww <public-rww@w3.org>
- Message-ID: <CAKaEYh+LRV_-9_qTE1WCUAMU5wm+svGhYg40i4kCHue6wvoUYw@mail.gmail.com>
On 2 September 2016 at 16:00, Kingsley Idehen <kidehen@openlinksw.com> wrote: > On 9/2/16 8:54 AM, Melvin Carvalho wrote: > > > > On 25 August 2016 at 15:36, Kingsley Idehen <kidehen@openlinksw.com> > wrote: > >> On 8/25/16 5:24 AM, Melvin Carvalho wrote: >> >> >> >> On 25 August 2016 at 04:10, Kingsley Idehen <kidehen@openlinksw.com> >> wrote: >> >>> On 8/24/16 2:00 PM, Melvin Carvalho wrote: >>> >>> >>> >>> On 24 August 2016 at 18:25, Kingsley Idehen <kidehen@openlinksw.com> >>> wrote: >>> >>>> On 8/24/16 9:08 AM, Melvin Carvalho wrote: >>>> >>>> >>>> >>>> On 24 August 2016 at 13:55, Kingsley Idehen <kidehen@openlinksw.com> >>>> wrote: >>>> >>>>> On 8/24/16 3:52 AM, Melvin Carvalho wrote: >>>>> >>>>> >>>>> >>>>> On 24 August 2016 at 04:17, Kingsley Idehen <kidehen@openlinksw.com> >>>>> wrote: >>>>> >>>>>> On 8/23/16 6:56 PM, Melvin Carvalho wrote: >>>>>> >>>>>> >>>>>> >>>>>> On 24 August 2016 at 00:28, Kingsley Idehen <kidehen@openlinksw.com> >>>>>> wrote: >>>>>> >>>>>>> On 8/23/16 5:36 PM, Melvin Carvalho wrote: >>>>>>> >>>>>>> yes, i was able to create a file, nice! >>>>>>> >>>>>>> On 23 August 2016 at 20:43, Kingsley Idehen <kidehen@openlinksw.com> >>>>>>> wrote: >>>>>>> >>>>>>>> On 8/23/16 2:25 PM, Melvin Carvalho wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On 22 August 2016 at 14:49, Kingsley Idehen <kidehen@openlinksw.com >>>>>>>> > wrote: >>>>>>>> >>>>>>>>> On 8/22/16 4:34 AM, Timothy Holborn wrote: >>>>>>>>> >>>>>>>>> Kingsley, >>>>>>>>> >>>>>>>>> Most of the interesting open data related platforms plug into >>>>>>>>> Virtuoso. >>>>>>>>> >>>>>>>>> >>>>>>>>> They support open standards. Virtuoso supports open standards. >>>>>>>>> >>>>>>>>> >>>>>>>>> I think you need to step it up a bit, and am happy to help, but am >>>>>>>>> unsure of the best way to go about it. >>>>>>>>> >>>>>>>>> >>>>>>>>> I am totally unsure of what Virtuoso has to add to this matter. >>>>>>>>> >>>>>>>>> >>>>>>>>> If SoLiD is Virtuoso compatible, I think the answer is bit of a >>>>>>>>> no-brainer. Question remains one of business >>>>>>>>> systems, rather than exclusively Tech. >>>>>>>>> >>>>>>>>> >>>>>>>>> Virtuoso supports all the open standards covered by SoLiD, and >>>>>>>>> some (e.g., WebID+TLS+Delegation). >>>>>>>>> >>>>>>>>> We need to speak clearly about these issues otherwise we have >>>>>>>>> nothing but confusion. >>>>>>>>> >>>>>>>> >>>>>>>> What will be really amazing is when Solid apps are tested to run on >>>>>>>> an openlink backend and vice versa. >>>>>>>>  >>>>>>>> >>>>>>>> Melvin, >>>>>>>> >>>>>>>> So why don't I share a folder endpoint [1] and the you try to use >>>>>>>> SoLiD to create a document in that folder? Naturally, I would need to grant >>>>>>>> access to you via your WebID (which I assume to be: >>>>>>>> https://melvincarvalho.com/#me) . >>>>>>>> >>>>>>>> Links: >>>>>>>> >>>>>>>> [1] http://kingsley.idehen.net/DAV/home/kidehen/Public/solid/ >>>>>>>> [2] https://kingsley.idehen.net/DAV/home/kidehen/Public/solid/ >>>>>>>> [3] http://kingsley.idehen.net/DAV/home/kidehen/Public/solid%2Cacl >>>>>>>> -- ACL doc (your webid has access to this too!) >>>>>>>> [4] https://linkeddata.uriburner.com/rdf-editor -- Editor that can >>>>>>>> be used to compare experience re. creation of document in the sample/qa >>>>>>>> folder. >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> Regards, >>>>>>>> >>>>>>>> Kingsley Idehen >>>>>>>> Founder & CEO >>>>>>>> OpenLink Software (Home Page: http://www.openlinksw.com) >>>>>>>> >>>>>>>> Medium Blog: https://medium.com/@kidehen >>>>>>>> Blogspot Blog: http://kidehen.blogspot.com >>>>>>>> Twitter Profile: https://twitter.com/kidehen >>>>>>>> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about >>>>>>>> LinkedIn Profile: http://www.linkedin.com/in/kidehen >>>>>>>> Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this >>>>>>>> >>>>>>>> -- You received this message because you are subscribed to the >>>>>>>> Google Groups "Business Of Linked Data (BOLD)" group. To unsubscribe from >>>>>>>> this group and stop receiving emails from it, send an email to >>>>>>>> business-of-linked-data-bold+unsubscribe@googlegroups.com. For >>>>>>>> more options, visit https://groups.google.com/d/optout. >>>>>>>> >>>>>>> -- You received this message because you are subscribed to the >>>>>>> Google Groups "Business Of Linked Data (BOLD)" group. To unsubscribe from >>>>>>> this group and stop receiving emails from it, send an email to >>>>>>> business-of-linked-data-bold+unsubscribe@googlegroups.com. For more >>>>>>> options, visit https://groups.google.com/d/optout. >>>>>>> >>>>>>> Melvin, >>>>>>> >>>>>>> Does that imply things are fine re. SoLiD or not? >>>>>>> >>>>>> One test is passing at least, which is a good sign! >>>>>> I think to say things are 'fine' we really need to develop a test >>>>>> suite and run tests. There may be other ways, but that >>>>>> seems to be tried and tested. >>>>>> >>>>>> Melvin, >>>>>> >>>>>> I am trying to avoid "OpenLink doesn't support SoLiD" cycles that >>>>>> keep on reoccurring. >>>>>> >>>>> Got it. But it requires testing and possibly some bug >>>>> fixing. >>>>>  >>>>> >>>>>> If there is a pattern that fails it should be identified and >>>>>> demonstrated. >>>>>> >>>>> This is where a test suite comes in handy. W3C working groups >>>>> typically require 1-3 years for this. I think we need a similar >>>>> process. There may be short cuts but it would normally require one >>>>> dedicated tester. >>>>> >>>>> W3C process != Practical Commercial process. >>>>> >>>>> Having worked on interop for more than 20+ years re., standards like >>>>> SQL, ODBC, JDBC, ADO.NET, HTTP, and others, the process has more to >>>>> do with willingness to collaborate than anything else. >>>>> >>>>> Given a server application (server) that implements standard X, there >>>>> should be N number of client application (client) users willing enough to >>>>> test interop as part of a practical QA process. Right now, the big issue is >>>>> that interop gets scoped to the wrong levels. >>>>> >>>> Presently I see people testing Solid against node-solid-server and >>>> gold. Previously I have seen testing against LDPHP. I've only seen you >>>> and sometimes me test against an openlink back end and that's when we have >>>> a bit of time free. >>>> >>>> Yes, but once again, its a case of understanding the roles of compliant >>>> servers and clients. Virtuoso is a compliant server. All you need is an >>>> endpoint and away you go. It either works or it fails. If it fails simply >>>> report what's failing. >>>> >>> Is virtuoso Solid compliant? Compliant to what? Has it been >>> tested? >>> >>> What do you mean by any of those questions? Put differently, why don't >>> you provide cURL based examples of what doesn't work, based on your >>> expectations? >>> >>> Does it handle globbing? >>> >>> cURL example please. >>> >>> Does it handle websockets? >>> >>> You now it does. >>> >>> Does it comply to the ACL spec? >>> >>> How did you end up creating a resource in a folder if it didn't comply >>> with ACLs scoped to your WebID? >>> >>> Does it support inboxes? >>> >>> What is an inbox? Put differently, how is it different from folder where >>> you store documents? >>> >>> Does it support Linked Data Notifications. >>> >>> No it doesn't . >>> >>> Does it comply to the sections of the latest solid spec? >>> >>> What are those? >>> >>> What browser coverage does it have, what breaks? These are questions >>> we are going through on a daily basis with other backends. >>> >>> Instead of asking these questions you could demonstrate your point with >>> a SoLiD client and/or curl interaction examples. >>> >>>  >>> >>>> What do I mean by "wrong levels" ? The fact that this kind of testing >>>>> gets lost in presumptive patterns rife with compilation and platform >>>>> dependencies e.g., open source and all the modules required to be located >>>>> and built. After that, testers then find out that they have to right code >>>>> to perform basic interop. >>>>> >>>> I think you mean people do not have the time to work though and fix >>>> bugs. >>>> >>>> No, I mean it is being approached the wrong way. What you need is: 1. >>>> List of compliant servers and their live endpoints 2. List of compliant >>>> clients 3. Folks testing the clients and the servers (this doesn't always >>>> have to be the developers of either client or server being tested). There >>>> isn't a single guideline that states: To verify or have some else verify >>>> SoLiD based interop, simply add your SoLiD compliant server and its >>>> endpoint to the list in the page at <some-server-usage-doc-location-uri> >>>> . To verify or have some else verify SoLiD based interop, simply add your >>>> SoLiD compliant client applications and a usage guide document link to the >>>> page at: <some-client-app-usage-doc-location-uri> . Post your results >>>> or share you experience via comments or reports to a document at: >>>> <some-interop-results-doc-location-uri> . >>>> >>> We are doing this constantly in the gitter channel. Behind that lies >>> the github solid repo which has active issue tracking. >>>  >>> >>>>  As it's a new technology inevitably there will be bugs, it needs a >>>> lot of persistence to work through. Openlink is not immune to bugs either, >>>> I have found and reported some myself. >>>> >>>> Do you have a link to SoLiD related bugs or issues? That's all we need. >>>> >>> Various repos under: https://github.com/solid >>> Pretty much all have issue tracking >>>  >>> >>>> Interop should simply be about compliant client N talking to compliant >>>>> server X. That's it. We don't need 6 months to pull that off, let alone 1-3 >>>>> years. >>>>> >>>>> I am happy to perform interop with anyone (partner or competitor or >>>>> customer) using the basic pattern outlined above. The end results are >>>>> mutually beneficial, as they should be, when working with standards >>>>> compliance. >>>>> >>>> Then just do it! >>>> >>>> I am confused. What is it that we haven't done? >>>> >>> Any kind of serious testing. My original point. If solid apps work >>> on virtuoso that's going to be a big win. Write a backend, write apps. >>> Test on virtuoso, test on node solid server, test on gold. That is the >>> test of compliance. Failing that, work on passing a test suite. >>>  >>> >>>>  I still believe the process we are using right now has not yielded >>>> fast progress, but a working group where people actually commit to >>>> deliverables does achieve interop. It's just a question of how much time >>>> each process takes. The thing about a WG is that you generally commit 1 day >>>> a week or as much as 0.5 of a FTE, per company involved. That's a more >>>> resource that is currently being employed. >>>> >>>> There is subtle confusion about my point reflected in your last two >>>> comments. If a SoLiD client fails to work with my Virtuoso instance, then >>>> simply indicate what the issue is. You can also share a list of SoLiD apps >>>> here and I can once again test them. That said, I have zero interest in >>>> compiling anyting or heading out on module graph bounties. I just want to >>>> install something and test. >>>> >>> Yes, I think we're talking high level perspective vs low level >>> perspective. The devil is in the detail. >>> I will be working on my own back end "solid live" and the acid test for >>> me will be whether solid apps can work with it. >>> >>> Your description of SoLiD, as exemplified by this exchange, isn't how >>> you make progress. First off, you need to be able to actually describe what >>> SoLiD actually is, clearly. Simply declaring things as compliant vs non >>> compliant, without any clarity isn't the way to generate uptake and interop >>> activity. What is the fundamental goal of SoLiD? What is does it actually >>> offer right now, that uniquely distinguishes it with regards to using HTTP, >>> WebDAV, LDP, Web ACLs, WebID+TLS, WebID+TLS+Delegation, SPARQL Graph >>> Protocol, SPARQL 1.1 etc. to perform Read-Write operations? Answering this >>> question is crucial :) Kingsley >>> >> Have you looked at this? https://github.com/solid/solid-spec >> >> Melvin, >> >> You know I've looked at that, and much more. We are having a public >> discussion and its really important that you (and other SoLiD supporters) >> embark on the following: >> >> 1. Articulate what problem SoLiD solves, uniquely >> >> 2. Demonstrate how SoLiD delivers on its value proposition via simple >> Client and Server implementations that just work i.e., no coding and >> compilation involved. >> >> We have a maze of technologies and "best practices" all conflated under >> SoLiD, unfortunately. That doesn't make for a sound interop basis when you >> have failure points at the following levels: >> >> [1] WebID Authentication using WebID+TLS protocol >> >> [2] WebID+TLS authentication protocol and Browser UX issue -- which is >> solved by WebID+TLS+Delegation protocol >> >> [3] Non-existent interop efforts across WebID+TLS, WebID+TLS+Delegation >> compliant clients and servers >> >> [4] Non-existent interop efforts across WebACL compliant clients and >> servers >> >> [5] All of the above for LDP compliant clients and servers; ditto SPARQL >> Graph Protocol and SPARQL 1.1 compliant clients and servers. >> >> Without 1-5 sorted out, you have nothing to work with, in a practical >> sense. >> > Melvin, What problem does SoLiD solve, uniquely? Put differently, you (and > SoLiD) supporters have to answer the following clearly: What is SoLiD? Why > is it important? How do I use it? > > My suggestion to test solid compliance might be to see if this profile > editor works with a given backend https://linkeddata.github.io/ > profile-editor/ > There's also an issue tracker linked to page > > Here's an illustration of what I experience: https://www.pinterest.com/ > kidehen/solid-interop/ . WebID+TLS handling is problematic, and that's > more to do with WebID+TLS interop issues associated with the chosen > WebID+TLS authentication module used in the framework to build that editor. > You've already successfully created resources in my Data Space where > authentication is WebID+TLS and/or WebID+TLS+Delegation based, leveraging a > WebACL. That's has nothing to do with SoLiD and everything to do with a > Read-Write leveraging existing open standards etc.. What is WebID+TLS? A an > open standards-based protocol for authenticating Agent identity, at > Web-scale . Why is it important? It enables Read-Write operations across > Data Spaces on an HTTP network using an HTTP user agent (or client). How do > I use it? Simply perform the following steps: 1. Make a profile document on > your local computer 2. Publish the profile document to an HTTP-accessible > location 3. Generate an X.509 Certificate using information from the > HTTP-accessible profile document -- during this process, use the HTTP > URL of the profile document + "#i" to create a WebID which is then used a > the value of the Subject Alternative Name field 4. Test WebID+TLS > authentication protocol using the product of 1-3. Tools: 1. > http://id.myopenlink.net/ods/webid_demo.html -- Example of a tool for > WebID+TLS protocol testing 2. http://id.myopenlink.net/youid/ -- Example > of a tool for X.509 Certificate endowed with WebID in SAN generation > (choose the WebID-Profile Document from the Profile data Provider > drop-down). > Thanks for trying it and taking the screenshots. Perhaps a next step could be to put those in a github issue, I wasnt able to click through without signing up to pintrest, which im not a member of. So I think this demonstrates that solid apps dont yet interoperate with openlink virtuoso. I think it's unhelpful to go into "blamestorming" mode but simply acknowledge this is the case, and that resources need to be allocated in order to achieve interop. The main issue is that everyone is light on resources, so that's why interop can take a while when doing things on a casual volunteer based workflow. Im not saying that a WG is the only way to achieve this, but rather, that the model of a WG ensures that participates allocate resources to achieving milestones. That makes the path to interop faster. I think 6 months is an optimistic timeframe (tho id love to see interop in that time), if we extrapolate back 6 months and see the incremental advances. > -- > Regards, > > Kingsley Idehen > Founder & CEO > OpenLink Software (Home Page: http://www.openlinksw.com) > > Medium Blog: https://medium.com/@kidehen > Blogspot Blog: http://kidehen.blogspot.com > Twitter Profile: https://twitter.com/kidehen > Google+ Profile: https://plus.google.com/+KingsleyIdehen/about > LinkedIn Profile: http://www.linkedin.com/in/kidehen > Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this > >
Received on Friday, 2 September 2016 14:14:36 UTC