W3C home > Mailing lists > Public > public-linked-json@w3.org > October 2013

Re: Pending issue of JSON-LD Implementation Report

From: Tristan King <tristan.king@gmail.com>
Date: Thu, 10 Oct 2013 17:40:03 +0200
Message-ID: <CACJ4DdKTTO3Ca01BTdh=aa3orXTB-Ww_VBTZRXQbyz4AOU_5vg@mail.gmail.com>
To: Gregg Kellogg <gregg@greggkellogg.net>
Cc: Peter Ansell <ansell.peter@gmail.com>, Linked JSON <public-linked-json@w3.org>
Here is the latest report (also available in other formats if you go to the
reports directory):


Everything passes except for the remote document tests, but i figured these
are less important than the core tests, and this is something I need some
extra time to think about how to tackle (which i don't have at the moment).

There's still some more work to be done before this branch can be merged
into master and released, hopefully it wont be too long before I have time
again to get this done.


On Tue, Oct 8, 2013 at 5:16 PM, Gregg Kellogg <gregg@greggkellogg.net>wrote:

> CCing public-linked-json@w3.org.
> We'd really like to have an implementation report for java-jsonld, and
> we're waiting another several days. If you can submit something with
> whatever coverage, later this week, that would be great!
> Gregg Kellogg
> gregg@greggkellogg.net
> On Oct 8, 2013, at 7:44 AM, Tristan King <tristan.king@gmail.com> wrote:
> I have had time the last few days to work on this and wanted to see how
> far I got by the end of today before I replied. I feel I'll need a few more
> days to get everything back up to speed and all the tests passing, so if
> you submit the report before then it's probably best to not include
> jsonld-java in it (or simply include the one Peter posted in an earlier
> mail). I'll respond again with an updated report when I'm done.
> Peter: I made a different branch because i've changed the code quite a bit
> and though it would be good to keep your branch to make it easier to
> regenerate the reports you did if we needed. In hindsight probably
> unnecessary, but it doesn't really matter in the end.
> On Tue, Oct 8, 2013 at 11:49 AM, Peter Ansell <ansell.peter@gmail.com>wrote:
>> I am not available this week for development and cannot regenerate the
>> report myself as i am only able to email from mobile. Tristan has started
>> to do some work on more required updates to reflect the spec changes but he
>> has created a separate branch to me so I am not sure what is happening at
>> this stage as I would have expected given this thread that he would have
>> added to the branch I created...
>> --
>> Peter
>> On 08/10/2013, at 4:01 PM, Gregg Kellogg <gregg@greggkellogg.net> wrote:
>> There's a fair chance we'll finalize the implementation report soon. If
>> Gould like Java-jsonld to be included, please let us know if and when
>> you'll be submitting a report, otherwise, we'll leave it out of the PR
>> implementation report.
>> Gregg Kellogg
>> Sent from my iPhone
>> On Oct 2, 2013, at 2:26 AM, Tristan King <tristan.king@gmail.com> wrote:
>> I actually started working on the updates a few weeks ago but work took
>> priority again pretty quickly. I have some time this week to do some more
>> work on it, will see how far I get by the end of the week.
>> On Wed, Oct 2, 2013 at 2:44 AM, Peter Ansell <ansell.peter@gmail.com>wrote:
>>> On 2 October 2013 09:40, Peter Ansell <ansell.peter@gmail.com> wrote:
>>> > On 2 October 2013 09:21, Gregg Kellogg <gregg@greggkellogg.net> wrote:
>>> >
>>> >> (By the way, Gavin Carothers was complaining about extremely poor
>>> performance of de-serializing large JSON-LD documents using java-jsonld on
>>> IRC the other day, not sure if you're aware of such a problem).
>>> >>
>>> >> Gregg
>>> >
>>> > I wasn't aware of that but we haven't performed any performance tests
>>> so far.
>>> >
>>> > I am aware of at least one place where we convert input documents into
>>> > a Java String, which is inefficient for very large documents and could
>>> > be improved to stream into Jackson using a Reader with a few minor
>>> > changes.
>>> I eliminated our conversion of all inputs to String's and pushed the
>>> changes to GitHub [1]. Now everything is streamed into Jackson from
>>> Readers (InputStreamReaders+UTF-8 for InputStreams).
>>> CC'ng Gavin to this so he is aware of that change which may improve
>>> his issues. We should now be streaming for both input and output, as
>>> long as code uses the non-String based methods from JsonUtils.
>>> Cheers,
>>> Peter
>>> [1] https://github.com/jsonld-java/jsonld-java
Received on Thursday, 10 October 2013 15:41:09 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:18:39 UTC