Thanks to Stephen, Ian and others for your further comments.  I would like
to add my voice to the suggestion that Adlib might provide some free XSLT
training, for example through a video, since we’ve found that techniques
that have worked perfectly well until now have unpredictably failed once
Adlib is transferred to SQL storage.  It would be a reassurance to be able
to see quickly some examples of what XSLT can achieve with Adlib so that we
would know how completely XSLT might be able to replace the adapl functions
that we have been using.

What I presume that XSLT won’t be able to do is to replicate FACS commands,
i.e. to query other databases and draw information from them into the
output.  I have spoken to colleagues (or, technically, ex-colleagues, since
I’m technically retired) in the Institute’s digital department today who
think the answer might be to combine XML output from two databases using
XQuery.  We’re going to look into how challenging that might be as a first
step.  I suppose it might even have the potential to solve the problem of
confining the list of authors to the ones that are in use on the exported
catalogue records without using the API, which we are not currently set up
to use.

As to production values of any videos – high production values might be
needed for videos giving an overview of Axiell products, that could be part
of the company’s “shop window”, but I agree that they aren’t necessary for
technical videos that will mostly be seen and used by existing customers.
If public perception is a worry, they could be accessible only through
MyAdlib, so that only existing customers could see them.

I still wonder why our adapls won’t run with our SQL database – or rather
they won’t run on the large number of records that we need to process,
although they work satisfactorily on small samples (which had lulled me into
a false sense of security until the end of last week).  I imagine it may
have to do with memory – I have seen RAM usage rise to 93% on our server
while one of the output adapls is running, although it doesn’t get anything
like as high on my laptop and the jobs still don’t produce the correct
output (and on my laptop they sometimes leave Adlwin in an unresponsive
condition – I’ve had Windows stop errors on my laptop immediately after
closing Adlwin (this is with version 7.4.17112.1 (x86) under Windows 10)).
Perhaps that’s a question for the help desk …

Thanks again for the contributions that have been made to the discussion so
far, and thanks in advance for any further ones,

Peter

Peter Salt
Senior Research Fellow
Institute of Historical Research
University of London | Senate House | Malet Street
LONDON  WC1E 7HU
t: +44 (0)1223 880493
f: +44 (0)1223 880493
e: [log in to unmask] 
Web: www.history.ac.uk
IHR Digital blog: blog.history.ac.uk


From: Adlib Users [mailto:[log in to unmask]] On Behalf Of Booth,
Fiona
Sent: Wednesday, November 1, 2017 11:03 AM
To: [log in to unmask]
Subject: Re: Adlib output adapls

Hi Ian,

I tend to use XSLT fairly sporadically to edit our online catalogue and I
often have to go back to the basics to do quite straightforward edits, so
videos would really help. 

I am aware there is more that XSLT can do than just the web catalogue so
even one demonstrating what XSLT can be used for with Adlib could be really
helpful. I wasn’t aware until talking to Rosie recently that some of the
work I was looking at getting written as an Adapl might also work in XSLT,
for example. I am sure there is more information online or within your
manuals but I personally learn more if I see something in action!  

I agree with Janice that a webinar could look at more in depth issues if
they couldn’t be broken into shorter chunks. The main thing for me is that
the content is delivered in a useful way and wouldn’t be too concerned if it
didn’t appear as a corporate video either.

Best wishes,
Fiona 

Fiona Booth
Digital Engagement Officer
Economic Development
Plymouth City Council
Plymouth Museums Galleries Archives
37 Tavistock Place
Plymouth
PL4 8AX
T +441752304762
E mailto:[log in to unmask]
http://www.plymhearts.org/


From: Adlib Users [mailto:[log in to unmask]] On Behalf Of Janice
Miller
Sent: 01 November 2017 10:11
To: mailto:[log in to unmask]
Subject: Re: Adlib output adapls

Hi Ian 

I think short 5 minute 'refresher' type videos would be incredibly useful.
Or have you considered webinars which would allow you to take people through
more complex issues and can be longer?  The National Archives (UK) have
produced some very good webinars which they record and make available later
on their website, for example
http://media.nationalarchives.gov.uk/index.php/webinar-using-discovery-natio
nal-archives-online-catalogue/ 

As have Preservica https://preservica.com/resources/webinars/ 

I don't think a high production value corporate type video would be very
useful myself. 

Thanks 
Janice 

Janice Miller, Archivist
Archives & Local Studies
East Dunbartonshire Leisure & Culture 
t. 0141 777 3142 

Search our online archive catalogue http://edlct.adlibhosting.com/






From:        Ian Rowson <mailto:[log in to unmask]> 
To:        mailto:[log in to unmask] 
Date:        01/11/2017 08:54 
Subject:        Re: Adlib output adapls 
Sent by:        Adlib Users <mailto:[log in to unmask]> 
________________________________________



Hi Stephen and everyone,

We are at the moment exploring a variety of possible future delivery methods
for training/documentation across all the Axiell product range, and video is
certainly one of the channels we are looking at.

I have to say we had been considering it more from the angle of basic user
training rather than topics which are more technical/specialised such as
this.

I'd be interested to hear feedback from the user community about which sort
of topics would be felt to be of most use for (short?) training/'how to'
videos? 

There is also a debate going on here that 4 or 5 mins  - or perhaps even
less - should be the absolute max run time, but then, some topics would need
longer than that?

Although on the face of it screen capture videos are not necessarily complex
to produce, another key debate here is about quality; If something is going
to be released as an 'official company video' on a topic, then there is an
argument that it needs to have good production values, and can't be just
something that a member of staff has quickly knocked together on their
desktop - but then that introduces barriers of cost/time to produce.

If you have thoughts on any of the above that I can feed in to our
discussions, please let me know...

Best regards

Ian


-----Original Message-----
From: Adlib Users [mailto:[log in to unmask]] On Behalf Of Stephen
McConnachie
Sent: 01 November 2017 07:51
To: mailto:[log in to unmask]
Subject: Re: Adlib output adapls

Hi Peter,

RE to use the API wouldn't we need to use Adlib web server? 
Yes, that's right, the API runs on the Microsoft web sever IIS (Internet
Information Services).

You could just export the XML from Adlib for Windows and manipulate it as
required with XSLT, but yes, XSLT skills required. I think your idea of
calling on technical skills within the university is a good one Peter, often
web developers or data science people have some XSLT skills.

Ian Rowson and David Forster: what about making an Adlib XSLT YouTube
training channel with some basic XSLT training, how to achieve simple but
useful things, to help kickstart some XSLT learning in the Adlib customer
community? I'd be happy to contribute.

All the best,
Stephen

Sent from my iPad

> On 31 Oct 2017, at 21:03, Peter Salt <mailto:[log in to unmask]>
wrote:
> 
> Forgive me, therefore, if this is a stupid question, but to use the 
> API wouldn't we need to use Adlib web server?  We do have a licence 
> for it, dating back to when we self-published with AHRC funding, but 
> we haven't operated a web server since the end of 2009 (or paid 
> maintenance for Adlib web server since then, for that matter!)
> 
> And (of course) we don't have a budget for the (re)training or 
> consultancy that might be needed, although we might be able to draw on 
> expertise elsewhere in the Institute of Historical Research (or the 
> University of London School of Advanced Study of which it forms a part).
> 
> Thanks again,
> 
> Peter
> 
>> -----Original Message-----
>> From: Adlib Users [mailto:[log in to unmask]] On Behalf Of 
>> Stephen McConnachie
>> Sent: Tuesday, October 31, 2017 6:02 PM
>> To: mailto:[log in to unmask]
>> Subject: Re: Adlib output adapls
>> 
>> Hi Peter,
>> 
>> I've no experience of using Adapl to achieve that kind of requirement 
>> as
> we do
>> what Ian's describing: we use XSLT to achieve pre-defined outputs, 
>> and for
> less
>> predictable requirements we call our Adlib REST API and retrieve XML 
>> or
> JSON
>> and implement processing on that XML or JSON once we have it.
>> 
>> From your example, I think you can definitely achieve your 
>> requirements in
> XSLT
>> Peter, with a bit of training in XSLT or a small budget to get a 
>> freelance
> coder or
>> Axiell to set your outputs up for you.
>> 
>> If you don't have training or coder budget, there are alternatives to
> explore, if
>> you have access to your Adlib API.
>> 
>> We have created a Windows command line tool that compiles a Python 
>> script which might help you to achieve it, although I can't remember 
>> if we
> hardcoded
>> our  an API details or allowed user to specify another API. It 
>> creates a
> pointer file
>> based on a given Adlib API query, and it can go further: create a 
>> pointer
> file in
>> People db, based on a query in Collect db. E.g. find me all Collect
> records where
>> author.person.name.lref > 0 (probably the wrong field name but you 
>> get the gist), then create me a pointer file in People db of all the 
>> persons in
> the author
>> fields.
>> 
>> Then you can output those People records with an XSLT in place to
> transform to
>> your publisher's required XML structure.
>> 
>> There are other ways to do it, I think, e.g. :  use the Excel 
>> =WEBSERVICE
> and
>> =FILTERXML to ask Collect via API, for all authors in every record 
>> with an
> Author.
>> Then use it again to ask for some more data per author. If you 
>> haven't
> used
>> those excel commands against your Adlib API I'd recommend having a 
>> go,
> they
>> offer a powerful method for using your Adlib API to ask complex 
>> questions
> and
>> deal complexly with the data you retrieve. They achieve what you can 
>> with
> VBA,
>> but they achieve it in formulae.
>> 
>> Easiest method is probably XSLT...
>> 
>> All the best,
>> Stephen
>> 
>> Sent from my iPad
>> 
>>>> On 31 Oct 2017, at 13:42, Peter Salt <mailto:[log in to unmask]>
wrote:
>>> 
>>> I have just made a very similar post to this one in the Axiell 
>>> LinkedIn group; I apologise to anyone who has already read it there.
>>> 
>>> As previously mentioned in this forum, we are trying to move our 
>>> Adlib database which has run with cbf storage since the late 1990s 
>>> to SQL
> storage.
>>> Our main catalogue contains around 650,000 records.  I am in the 
>>> final stages of testing and I am finding that output adapls using 
>>> FACS commands are not reliable when dealing with large numbers of
records.
>>> (They also seem much slower when running on the SQL version than on 
>>> cbf, as I've already found with other tasks such as importing and 
>>> re-indexing, but this is a secondary consideration.)
>>> 
>>> For example, we use an output adapl and print the output to file in 
>>> order to generate XML in the exact format requested by our publishers.
>>> The adapl that produces our file of authors uses a FACS "read" 
>>> command to check the catalogue to see if the author name is in use 
>>> there, and omits it from the output if it is not in use.  With cbf 
>>> storage I have been able, since 2009, to generate three XML files of 
>>> authors, each containing about 100,000 records, and checks done on 
>>> each file suggested that the process was completely reliable, 
>>> including all author records that should have been present and omitting
the others.
>>> With SQL storage, I have managed to produce a satisfactory file of 
>>> about 10,000 author records, but for anything larger than that, the 
>>> job will apparently complete but, on inspection, is found to have 
>>> omitted authors that should have been present, although on occasion
> Adlwin
>> crashes while the job is running.
>>> 
>>> Is this a known limitation of Adlib with SQL storage?  Do we need to 
>>> find a different way of generating this data (Ian Rowson of Axiell 
>>> UK mentioned to me in a different context that "these days" the 
>>> company "tend to use XSLT for creating custom output formats where 
>>> possible" - but can we replicate FACS functions in XSLT)?  Or should 
>>> we be looking for something in our adapls that, while not a problem 
>>> with cbf storage, becomes a problem with SQL storage?
>>> 
>>> Peter Salt
>>> Senior Research Fellow
>>> Institute of Historical Research
>>> University of London | Senate House | Malet Street LONDON  WC1E 7HU
>>> e: mailto:[log in to unmask]
>>> Web: www.history.ac.uk
>>> IHR Digital blog: blog.history.ac.uk
_________________________________________________
Scanned by Serendipity Interactive - MYmailSAFE


****************************************************************************
*******
DISCLAIMER:
This email and any files transmitted with it are intended for the use of the
individual or entity to whom they are addressed. It may contain information
of a confidential or privileged nature.
If you have received this email in error please notify the originator of the
message and destroy the e-mail.
East Dunbartonshire Council can not be held responsible for viruses,
therefore please scan all attachments.
The views expressed in this message are those of the sender and do not
necessarily reflect those of East Dunbartonshire Council who will not
necessarily be bound by its contents.
****************************************************************************
*******
http://www.eastdunbarton.gov.uk

________________________________________
IMPORTANT: This e-mail (including any attachments to it) is strictly
confidential and intended solely for the person or organisation to whom it
is addressed. It may contain privileged, confidential or sensitive
information. If you are not the intended recipient, you must not copy or
distribute it to any other person or take any action in reliance. If you
have received it in error, please notify your system manager and the sender
as soon as possible and then delete it from your system.