Emerging Functional Requirements for a Search & Discovery TBM Application Profile to support post-production material
This document is intended to be delivered and read in conjunction with the TBMAP Report which provides relevant qualification, background and status information regarding the material presented here. The key points of this document allowed the summary set of Functional Requirements presented in Appendix 4b to be derived.
Functional Requirement(concept information) / Example Use Cases
‘In my role as x, I want to be able to…’ / Initial Notes & Questions opened for comment
To reflect the structural nature of the media
i.e. express the relationships between "parts" and the "wholes" from which they were derived, including the start- and
end-point(s) of those part(s) within the whole.
Note: This requirement covers concepts identifying the relationship of the physical ‘whole’ to an equivalent digital ‘whole’ with subsequent identification and retrieval of associated ‘segment or clip’ ‘extracted still or frame’ items and frame by frame (sequential and non-linearly spaced) or section by section descriptions and revisions. / - Identify and make the connection between post-edit items that are related to each other on the basis they each include all or part of a defined work.
- Retrieve all assets and asset types associated with a ‘whole’ as a coherent set.
-Easily differentiate between the assets using a consistent information set tailored to each part.
- Easily know if the material is what I expect it to be (e.g. using visual recognition pointers in returned records such as clips from the whole or thumbnails)
- Obtain access to individual segments or clips complete with their associated time-based metadata (e.g. shot-by-shot descriptions).
- Obtain information about the time-based relationship between a segment or clip and its ‘whole’ asset, including:
in- or start-point in reference to whole,
out- or end-point in reference to whole,
segment duration, and associated labelling (segment title, segment keywords, relation to whole clip from which it was taken) / -Frames are being treated as ‘very short parts’ of a ‘whole’.
-Describing and searching detail within individual Frames or extracted Stills is assumed to be able to defer to and be handled by ‘image’ metadata rather than re-create this functionality within ‘time-based media’ profiles.
-‘Frame scraping’ is also assumed to provide information that is best handled as ‘image’.
- It is assumed that a ‘still’ and sets of extracted stills can be used in a variety of ways to illustrate search results e.g. as individual thumbnails or as sub-image collections which may or may not be presented sequentially or played back in a time-based way. This AP only seeks to be able to identify extracted frames or stills back to their related available ‘whole’ or ‘segment’ items and not to describe additional functions that may be built around extracted material.
-It is assumed that only ‘whole’ and ‘segment’ will be supported for audio.
-Where audio supports visual time-based media descriptions or captions may reflect associated visual material.
To accommodate the technical nature of the media
e.g. production information, conversion or compression support, media-specific metadata standards, pre-embedded and/or automatically generated file metadata. viewing requirements. / For all of the different whole and parts returned from a search:
-Select from the different bandwidths (e.g. for streaming), download formats, media types (e.g. DVD purchase) and qualities available for the same piece of material and know which will actually be delivered in what format by what route.
-Obtain URL (or PURL) for the stream representing each delivery quality (e.g. to embed it in my course or presentation material).
-Identify related material by technical or production criteria (e.g. format, see attributes list)
-Know the Running Time (Duration)
-Know what metadata might be contained once I have my own copy of the file (e.g. that can be grabbed from files where appropriate/technically feasible)
-Know what technical requirements I need to meet to view the material, including if the codec is different from the filetype. / -Users may wish to search for tools or functionality that will support them in edit tasks using desktop tools to promote content re-use in sharing environments. Or perhaps to see if material already exists so they do not have to create it. There is some overlap in technical and search requirements to meet this need.
- However, we concur with the approach taken in recommending a set of preservation metadata presented in AHDS ‘Digital Moving Images and Sound Archiving Study’ report :
‘An approach based on format-specific standards would, most likely, be unsustainable and hinder interoperability in view of the number of formats to be found…… The resulting element set should be regarded as a bare minimum. Most repositories should record richer metadata than it mandates by adding elements from other recognised standards to form application profiles suited to their specific needs. Moreover, the element set makes no assumption as to the way the metadata is stored or contained (e.g. MPEG-21 or METS).’’
And that there may need to be augmentation from other recognised standards with format-specific elements added only where omission would appear likely to compromise the overall functionality of the profile.
To support common delivery mechanisms
e.g. streaming media, download, physical media / - Have the material supplied by streaming and know whether or not it is only available by streaming and also that URLs provided are stable (e.g. for course use)
- Have access to a download copy and be advised if there is no download copy available or users cannot obtain a local copy (i.e. if online or offline only and local backup versions are available).
- Where possible have direct access to the complete functional resource
- Have access to preview versions where these are available
To support user-required tools
e.g. viewers, technical processes, editing systems, identification of pre-entered metadata provided via automatic sources, media manipulation / - What tools are available to allow me to view the resource?
- Know where the metadata I am viewing comes from (e.g. my university supplied this part of it, the tool supplied that part of it, it came with the resource)?
- Find tools to manipulate video
- Enter routine metadata once and have the system learn my patterns for different jobs
- Search by:
number of times that clips have been downloaded FE/HE users (use statistics),
Amazon-style reviews and rankings. / Although this user requirement frequently occurs in media projects, it is considered out-of-scope for this profile. Effectively the user requirements describe functionality and interfaces built for the user based on the metadata that is supplied.
This includes where users may wish to search for tools or functionality that will support them in edit tasks using desktop tools to promote content re-use in sharing environments.
The profile should provide enough technical metadata to support TBM resource search & discovery but leave the presentation application or the client to make the association with tools that are required.
If the tool itself is deposited as a resource to be discovered e.g. as part of a thesis, it is questionable as to which profile would best support that tool, although basic metadata can still be used.
To assist with identification and retrieval of accessible versions and supporting material
e.g. subtitles, sign language editions, language variations, audio-only editions, soundtracks, text transcripts, title changes, translations, scripts, educational resources, play lists, tutor or course notes and guidance, aggregated resources, additional subject-specific information not included in the profile (references to external sources of), posters, trailers, programme guides. / - Identify and make the connection between the post-edit items that are related to each other on the basis they each include all or part of a defined work, and all available non-TBM material that is designed to support that work.
-Clearly indicate which items relate to which whole, clip or frame.
- Retrieve all associated main and supporting assets and asset types as a coherent set.
-Easily differentiate between the main and supporting assets using a consistent information set.
-Locate existence of alternate forms of the whole and/or its segments that are able to support a variety of disabilities and indicate this to be the case.
-Inform of any viewing conditions and restrictions including ensuring any confidentiality clauses are clearly stated and that enforcement is supported where necessary. / Related material can be part of the production and distribution process, or generated for genre specific users to provide them with tailored information and services e.g. pedagogical support information or access versions.
Note that one of the BUFVC databases implements an advanced search which includes what is essentially a vocabulary for technical aspects of Television which also includes accessibility information such as subtitling, sign language as well as technical accessibility (e.g. widescreen or ceefax etc.).
It may be a requirement for the profile to be able to accommodate several specialist vocabularies to support different user needs e.g. academic subjects, television with access metadata.
To allow navigation between and cross-identification of all associated structures and materials
e.g. support for providing unique identifiers that can cross-reference. / -Be able to navigate between different versions representing the same original work particularly where These are only technologically distinct
- See association of items that relate to or are derived from the primary content source e.g. stills, segments, transcripts, scripts, posters, trailers, academic supporting material.
- Be notified of related works to the same content / The project is assuming that identifiers will be generated by the metadata providers and the systems that generate metadata on their behalf, with the role of the profile being to specify the entities that require identification.
Notification is probably a function that can be built based on metadata.
[It is anticipated that the main search/browse navigations requirement would come out of the modelling exercise.]
To provide basic content descriptor information and compatibility with additional search entry points
e.g. basic collection-oriented information or identifying agents. Possibly inclusion of properties required to fulfil other requirements such as describing provenance, institution of origin, research funding body, and enabling national and regional views of repository metadata. / -View written descriptions of particular items
-Be able to freetext search
-Search for content (support and TBM) by topic or collection
- Allow streamed resources to be logged as part of a collection where they are not directly held (e.g. to support wider access)
-See supply and provenance information
-Know age of source material [used in a digital copy] as a potential indicator of degradation of the content.
- Find information on programme type,
sponsor, commissioning department,
target audience. / The profile is assumed to be a modular structure where curators, repository managers and developers are free to extend and/or build around the profile to accommodate specialist requirements (e.g. DC-Education, Preservation). However, this requirement is included to allow the profile to also be self-contained, provide wider exposure for TBM material held apart from systems hosting the metadata, and to offer compatibility with existing JISC services such as INTUTE and the recommendations made for preservation of TBM by AHDS. The attributes required or recommended are indicated in attributes reference.
A range of attributes seem to be important in HE, many may be attributed to the large range of specialist uses. These are noted in the attributes document and can be consulted from that basis.
Use information including availability, rights, restrictions, obtaining (changed purpose, changed conditions, new user group), expiry and related preservation conditions
e.g. licence terms, validity periods, enforcement of terms such as retiring content at end of paid period, availability (e.g. of archived physical material), user group or viewing restrictions, and information on how to obtain the material for alternate purposes, associated agents (e.g. copyright holders, sales or rights clearance points). / -Have Access to Rights & Use information
-Be provided with the expiry date for any licence agreement
- See clear re-use or re-development information
-To be provided with terms of use and rights information for each asset associated with a primary content source.
-Have open access or special conditions around materials highlighted
-Filter information that is not relevant to me (e.g. if different prices occur for different organisations)
- Know cost and conditions to me in different roles
- If only demo material is available from a search, show me how to get the real thing e.g. link from demo clip through to details of supply and purchase.
- Have availability information for the product
-See information about the supplying organisation (e.g. radio station or network library), Description, Contact details (agent: name, position, address, phone, fax email web) / ‘Availability’ can also provide information on the status of a physical resource that is recorded Relevant for e.g. material that has to be retrieved from an archive, particularly where records are primarily for research data on resource existence and use as opposed to availability. For example, was available, unknown availability, poor condition, restored, obtaining a copy is possible, last use, etc. For example where an original physical source has to be located from which to make a supply copy. May also refer to whether a copy is on loan from a library or available for loan.
Should preservation be treated as a ‘genre’ i.e. specialist application area that may or may not be included for compatibility – if it isn’t then show how the profile may accommodate it.
Profile should just provide basic hooks and pointers onto which a rights description language can be supported. Out-of-scope to develop, choose or recommend a rights description language.
Is it good or bad practice to combine credits for video/audio as rights statement to indicate copyright holder (c. so and so)??
Extensibility for tailoring to genre-specific requirements and added-value services
e.g. to support subject-specific material such as Life Sciences information, industrial genres such as film or music industry metadata, course and education metadata. / -Access to Keywords and user tagging
-Support for Multiple Controlled Vocabularies / It may be a requirement for the profile to be able to accommodate several specialist vocabularies to support different user needs and indicate how content is used in different situations, such as:
- academic subjects (e.g. MeSH, LCSH, RDN/LTSN resource type vocabulary, LTSN pedagogic terms vocabulary),
- curricular elements,
- television with access metadata,
- programme type,
- channel.
- name authority,
- date/time, place (GPS) link to geospatial
This can be combined with user support sources requirement to support e.g. blog or resource pointers where users extend the support information relevant to their groups.
Possibly allow for some sort of general "annotation" or "review" or "use note" or whatever.
Metadata can also be records for material stored elsewhere e.g. media server.
This requirement also encompasses support concepts such as:
- attract and maintain a database of expressions of interest ….so that when teachers have expressed an interest in a video currently unavailable in a digital format, they are informed when it becomes available.
-email link to a friend
-Analysis of search information e.g.(a) which programmes have been downloaded by which departments in which HE and FE institutions (b) which programmes are being used (c) which programmes, although downloaded, are not being used (d) how users discovered the programmes which they used (e) what other topics users need to have available.
Handling of complex objects
-Handling material embedded in other representations e.g. within or as part of other documents, such as word or PDF.
- Handling material that is one item but separated into granular components for storage e.g. thesis with code or TBM ‘appendices’.
-Establishing an understanding of existing work in this area and prioritising requirements, including e.g. for packaging and container standards such as SMIL, METS or MPEG-21.
- Decisions on how to include time based media in the modelling of complex objects. / -retrieve and/or re-use podcasts / Podcast material falls into this category; a media file with an RSS component to enable searching. Should these be stored together as an object and handled, or should the profile be able to accommodate basic elements commonly encased in the RSS? Should both files be stored independently and the RSS file treated as related material? Can SMIL just be entered as a format?
1