[NLPL Task Force (A)] draft version of updated NLPL software installation guide

Martin Matthiesen martin.matthiesen at csc.fi
Wed Nov 8 11:00:57 UTC 2017


Hello all, 

I read the guide and I think is is very clear, thanks Stephan! (And sorry for the latency...) 

I comment a bit on Jörg's comments below: 

> From: "Jörg Tiedemann" <jorg.tiedemann at helsinki.fi>
> To: "Stephan Oepen" <oe at ifi.uio.no>
> Cc: "infrastructure" <infrastructure at nlpl.eu>
> Sent: Tuesday, 7 November, 2017 23:50:43
> Subject: Re: [NLPL Task Force (A)] draft version of updated NLPL software
> installation guide

> Hi,

> Thanks, Stephan for migrating the guide to the Wiki. This is very useful.
> I quickly read though the webpage and I have no immediate suggestion. I guess
> that the recommendation about language codes is just a recommendation and that
> we can keep, for example, two-letter codes in case this is established for a
> certain resource. For example, in OPUS I use them and it would mean a major
> change in certain scripts to move to 3-letter codes (even though I would
> probably like that)

> About the question with module names: It is correct that I suggested the prefix
> nlpl at some point and I used that in my modules so far. I don’t know if it
> make sense and no strong opinion from my side either. My idea was to make a
> distinction between nlpl-specific modules and the ones that made it into the
> general repository of module files. It could also flag a bit our activities to
> make people aware of NLPL.

> The downside is that writing the nlpl prefix can be annoying and does not add
> much of a value.
> But sometimes it makes sense to use this kind of branding. And especially
> because NLPL is not very established among our colleagues this could actually
> work to make people aware of our activities?! What’s your opinion?

I think that NLPL could be mentioned in the help or module description. I think brandind and awareness is important, but for the end user the package counts, not so much where it came from. Imagine we would have been the first a while ago to install "R" as "nlpl-R". Everyone now would wonder how that is different from the "R"s around. Only to find out it is not. Tools wandering from us to general use would be a great thing, I think. 

> Otherwise, for me it’s fine to announce the link.
> We also started the software catalogue: [
> http://wiki.nlpl.eu/index.php/Infrastructure/software/catalogue |
> http://wiki.nlpl.eu/index.php/Infrastructure/software/catalogue ]
> Would that work as a starting point?

> A quick question about the placement of documentation of the OPUS activity on
> the wiki: I thought about placing it under “large corpora” even though this is
> actually another activity. But in the long run it probably makes sense to have
> it as a category below corpora and not on its own. Does that make sense?

I would say yes. I think we should keep the end user in mind who would not want to understand our project plan to find documentation. 

A question from the page: 

    * Handling of data sets with access restrictions 

Did we not at some point decide we simply don't deal with restricted data? I know that some colleagues of mine are working on a solution that uses Shibboleth, OAuth and encrypted mounts, would we want to pilot such a solution? I already volunteered for Kielipankki, but have not heard back yet. 

Looking forward to meeting you all tomorrow! 

Martin 

> By the way, I thought about syncing at least the core XML files from OPUS to
> abel. Should I do that and how much space am I allowed to fill in that case? I
> can skip all the derived data sets like bitexts in different formats, word
> alignments, parsed data etc …

> See you soon in Oslo!
> Jörg

> **********************************************************************************************
> Jörg Tiedemann
> Department of Modern Languages [ http://blogs.helsinki.fi/tiedeman/ |
> http://blogs.helsinki.fi/tiedeman/ ]
> University of Helsinki [ http://blogs.helsinki.fi/language-technology/ |
> http://blogs.helsinki.fi/language-technology/ ]

>> On 07 Nov 2017, at 12:44, Stephan Oepen < [ mailto:oe at ifi.uio.no | oe at ifi.uio.no
>> ] > wrote:

>> colleagues,

>> the NLPL wiki was down until mid-day yesterday :-(. i am talking to
>> USIT now about avoiding longish downtimes in the future and hosting
>> the wiki (and potentially other NLPL services) in a manner that makes
>> it easier to migrate transparently to a replacement host, if need be.

>> either way, i have now completed migration of the software
>> installation guide to the NLPL wiki:

>> [ http://wiki.nlpl.eu/index.php/Infrastructure/installation/guide |
>> http://wiki.nlpl.eu/index.php/Infrastructure/installation/guide ]

>> i would like to announce the link to the team sometime tonight, if
>> possible. martin (and others), would you have a chance to quickly
>> proof-read the updated version?

>> i noted one missing constraints which i thought we had decided to
>> enforce at the kick-off, viz. that all NLPL module definitions be
>> prefixed with ‘nlpl-’, .e.g. ‘nlpl-moses’ for the Moses software. i
>> believe there may have been some follow-up discussion on slack, which
>> i may have missed, and yves so far has omitted the NLPL prefix.

>> i have no strong feelings either way; does anyone? joerg, you had
>> originally made the prefixing proposal, i believe.

>> cheers, oe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nlpl.eu/archives/infrastructure/attachments/20171108/52e6ee76/attachment.htm>


More information about the infrastructure mailing list