On Tue, Jan 29, 2013 at 7:48 AM, Andy Seaborne <[email protected]> wrote:
> On 29/01/13 10:57, Rob Vesse wrote:
>>
>>
>> On 1/29/13 9:28 AM, "Andy Seaborne" <[email protected]> wrote:
>>
>>> The prefixes improvement are good timing for me because some new
>>> possible writers in
>>>
>>> https://github.com/afs/riot-output
>>>
>>> are making progress.  Writers done, working on the registry and
>>> extensibility.
>>>
>>> (moved out of /Experimental so as not to cause many small commits to
>>> clog things up - I'm just as happy to use svn/Experimental is you want
>>> to see the fiddling about before first usable release)
>>
>>
>> I assume GitHub has the notion of email notifications so you could turn on
>> notifications of your git commits on that repo and have them sent to the
>> [email protected] list like the SVN notifications do?
>>
>> That would be nice as it allows us all to see and comment on work that
>> goes on even if it doesn't happen in the SVN repo
>
>
> If we are agreed on seeing the noise, I can move it back to ASF.  I'm
> neutral as to location.
>
> I would appreciate review and improvement - feel free to just edit the code
> for the "doh!" things.  While review should be able learning better ways, it
> is also about finding the stupidities, which are often more numerous.
>
>

I would slightly prefer ASF, so that I can see the commit messages.  I
do try to read/skim most of them, if only to see what bits people are
working on.


>>
>>>
>>> It includes pretty TriG writing, pretty Turtle writing, N-triples and
>>> N-Quads that should be faster (less string bashing, more buffering).
>>>
>>> Faster URI to prefix name will be valuable.  I was going to have to
>>> worry about it and now I don't :-)
>>
>>
>> Btw I have renamed FastPrefixMap to FastAbbreviatingPrefixMap to better
>> reflect its intended usage and capability.
>>
>>>
>>> There is a PrintStreamRDF that does streaming output of Turtle (no
>>> lists) and makes something easier to read than N-triples.  I was going
>>> to add it to the riot parser commands as an output option.  This will be
>>> performance critical.
>>
>>
>> This sounds really good, definitely something I will be using
>>
>> Rob
>>
>>>
>>>         Andy
>>
>>
>

Reply via email to