QA For Metadata: Exercise 1

1
QA For MetadataExercise 1
This presentation reviews the exercise in use of
the ltacronymgt and ltabbrgt HTML tags which
highlight general issues concerning use of
metadata

• Brian Kelly and Amanda Closier
• UKOLN
• Gareth Knight
• AHDS

2
QA For Metadata Exercise 1

• In small groups you attempted exercise 1a
• Policies to ensure your acronyms and
  abbreviations are interoperable
• Procedures to ensure your policies are
  implemented correctly
• We will now review
• Areas of difficulties in using these tags
• Justification for using such metadata
• Interoperability issues
• Quality Assurance

3
About ltAcronymgt and ltAbbrgt Tags
Background

• The ltabbrgt tag indicates an abbreviated form
  (e.g., WWW, HTML, URI, et al. etc.) and includes
  initialisms.
• The ltacronymgt tag indicates an acronym (e.g.,
  FAIR, CETIS, etc.).
• The title attribute can be used to provide the
  full or expanded form of an expression.
• Examples
• ltabbr title"World Wide Web"gtWWWlt/abbrgt
• ltacronym title"Joint Information Systems
  Committee"gtJISClt/abbrgt

See W3C's definition of these tags
4
Benefits
Background

• Accessibility
• Speaking browsers will speak out
• Individual letters of abbreviations e.g. WWW as
  Double-You Double-You Double-You Double-You
• Acronyms as a word e.g. JISC as Jisc
• Interoperability
• Tom Heath's acronym robot can create an automated
  glossary
• See acronym tool - lthttp//www.materials.ac.uk/acr
  onyms/gt

5
Acronym Tag Issues (1)
Issues

• People dont know this tag exists!
• Confusion over whether ltacronymgt or ltabbrgt is
  used
• All acronyms are abbreviations, but all
  abbreviations are not acronyms
• Acronyms can be considered a subset of
  abbreviations
• Lack of consistency in way words are pronounced
  e.g. FAQ, SQL, URL,
• Changes over time e.g. origins of radar, laser,
  etc.
• Cultural differences (US vs UK English)

See discussion of issues
6
Acronym Tag Issues (2)
Issues

• Some abbreviations are confusing because they
• Are excepted into everyday language e.g. info,
  Mac
• Are abbreviated in one language but spoken in
  others e.g. e.g short for exempli gratia but used
  as for example
• No longer mean anything e.g. UKOLN
• Should they be marked up? Does the reader need
  more information? How relevant are they? Do we
  use
• ltabbr title"exempli gratia" lang"la"gte.g.lt/abbrgt
  
• ltabbr title"for example" lang"en"gte.g.lt/abbrgt
• ltabbr title"for example"gte.g.lt/abbrgt
• e.g.

7
Acronym Tag Issues (3)
Issues

• Issues about how the terms are marked-up
• Nesting decisions e.g. FAQs in the tags vs just
  FAQ with the 's' left outside (FAQs)
• Capitalisation in the meanings e.g.
  hewlett-packard vs Hewlett-Packard
• Punctuation e.g. I.T. vs IT
• Formal expansion of chatty text
• Changes in meaning of acronym

ltacronym title"Facilitating Access to
Information on Learning Technology for
Engineers"gtFAILTElt/acronymgt or ltacronym
title"FAILTE stands for Facilitating Access to
Information on Learning Technology for Engineers.
Failte is also the Gaelic word for Welcome and is
pronounced fawl-sha"gtFAILTElt/acronymgt
8
Acronym Tag Issues (4)
Issues

• Do we markup phrases based on
• Policies definitions
• Browser support
• Note that Opera Mozilla support the tags but IE
  does not

• Issues
• The markup takes time and as the most popular
  browser doesn't support it, it's not worth doing
• It's a standard, so we should do it
• It provides interoperability, so we should do it
• IE is evil, so we should do it

9
Acronym Tag Issues (5)
Issues

• Markup errors
• ltacronym alt"foo"gt rather than ltacronym
  title"foo"gt
• ltabbrevgt rather than ltabbrgt
• Markup in attributes
• Use of ltacronym alt"ltbgtfoolt/bgt"gt or ltabbr
  alt"ltugtWlt/ugtorld ltugtWlt/ugtide ..
• Invalid characters
• Unescaped character entities such as (amp)
• Incorrect content
• ltacronym title"Extended Markup
  Langauge"gtXMLlt/acronymgt

10
Acronym Tag Issues (6)
Issues

• How should you create and manage your acronym and
  abbreviation markup?
• Create by hand
• Functionality provided by your CMS
• Dedicated tools e.g. acrobot

http//www.accessify.com/tools-and-wizards/acrobo
t/
11
Acronym Tag Issues (7)
Issues
http//www.materials.ac.uk/acronyms/location.asp

• Can the benefits provided justify the costs of
  implementation?
• Automated Glossary
• The acronym harvester and glossary tool provide a
  lightweight mechanism for producing a glossary
• If every JISC project marked up acronyms on their
  home page (project names, organisations,
  technologies) this could provide a simple but
  effective mechanism for providing a glossary

Note the acronyms have been marked up in QA Focus
documents project names come from the case
studies
12
Acronym Tag Issues (8)
Issues

• Choosing Or Creating A Schema
• What schema should we use for our metadata (i.e.
  how do we structure our metadata)?
• Do we use a standard schema (good for
  interoperability) or develop our own (may provide
  better support for local needs)
• Acronym Example
• It would be useful to split acronyms into project
  names, organisations, technologies and other
• Could be implemented with ltacronym class"org"
  title"Joint " gtJISClt/ acronymgt
• But how do we get consensus on schema, implement
  support in tools, validate, get buy-in,

13
Acronym Tag Solutions
Solutions

• To deal with the issues when using the acronym
  and abbr tags QA Focus have developed
• A documented policy
• Oxford ED ? No punctuation
• Formal definition additional info in normal
  text
• A set of procedures
• Staff development ? Automated validation
• Ad hoc manual checking to spot content errors
• Justification automated glossary for Web site
  (possibly contributing to glossary across
  projects?)

Ideally support would be built into a CMS, but we
currently don't use a CMS
14
Conclusions

• This simple example illustrates several points
• Metadata is not just about resource discovery
• Metadata needs managing
• Before you can manage your metadata you will need
  policies so you (and others) have an agreed and
  shared understanding
• It is always useful to make use of a standard
• But standards can sometimes be flawed,
  inconsistent, etc.
• Support for your metadata may also be incomplete
• You should think carefully about your approach
  for managing your metadata
• You don't have to use metadata!

15
Exercise 1b

• You wish to create and manage metadata for your
  5,000 tracks on your 20 Gb MP3 player. Additional
  challenges application hardwired in player, no
  open source solution
• Issues
• Choice of file format Universal MP3 or better
  but more proprietary WMA format
• Selection of genres Leave to central database or
  use own schema e.g. house, acid, garage vs modern
  rubbish -)
• File names Player plays tracks in alphabetic
  order so need artist track_no. track_title
  and not track_title. But if multiple artists on
  CD need CD_name track_no. track_title
• Interoperability Decisions taken for me my
  player or allow for further players, family's
  music metadata,
• Other issues Compilation CDs, collections,
  physical CDs, ...

16
Conclusions (1b)

• Further conclusions
• Policies are needed even for small-scale personal
  applications
• You can't always program your way out of
  difficulties
• There may be conflicts between local usage and
  wider interoperability
• There is a need to be aware of how applications
  will use your metadata, so you shouldn't develop
  your metadata model and policies independently of
  the applications

Note that managing a 20 Gb MP3 music player
containing 5,000 tracks has similarities to
managing a small library!
17
Any Questions?