Next Gen Catalog Requirements
Intro/Background
We need a short intro here, plus some instructions for those adding to this page. The intro should define the scope of the page (not of the NGC)
Requirements (cooked)
This area is for edited requirements. See this as a single, coherent document, without (too much) repetition
Data and database
- Expanded content Meaning reviews, tables of contents, full text.
Search/Retrieval/Display
- Metasearch Search potentially includes all resources available to the user.
- Relevance ranked output This is a commonly desired feature, but it is only possible if the data are available to create such a ranking. So our discussion will need to focus on how ranking could be done, not whether it should be done.
- Faceted navigation Can include: general discipline; level (basic, academic); audience; genre; date of creation; topical time frame; obtain options (online, on shelf, purchase);
- Spell-checking, stemming, etc. Directly related to keyword searching, these user aids are language-dependent and may have to be related to the "language of the catalog."
- Full Unicode character sets for search and display ñμƑƑ ƧäÎÐ ☺
- Search inside the book (where full text is available)
- Display includes images Covers, maybe also title pages and tables of contents
- Brief display is topical Reveals what resource is about.
- Brief display is work-related A merged set of variations on the same work
- Readers advisory What should I read next? What would be a good starter book on this subject? Would I like this particular book?
User functions
- Selective sort capabilities The ability for users to sort and re-sort retrievals by a variety of keys. The investigation of this feature should include the desire to sort by data elements that occur more than once in the record (e.g. subjects). See: Yee article.
- Personalization A broad category, but we can start with MyLibrary and move on to other features. Some features are:
- Ability to create personal or shared bibliographies and booklists, with tagging
- User profiles. Allow users to set their own start page, default search types, default displays, etc.
- Notification Services Receive online or email or RSS notification of due dates, holds available, new items received (by topic area)
Staff functions
- Optional layer of abstraction The cataloger interface should be able to translate the codes used in the catalog record into the plain language of the cataloger, if that is how the cataloger wants to see the record while editing it. (Example: "eng" from a MARC record could become "English", "Anglais", "Ingles", etc.)
System functions and features
- Service-layer approach System is not monolithic, but contains parts that are brought together by one or more service layers and user interfaces.
- Exposure of contents via open API's A goal of the catalog needs to be to interact with other information systems, and therefore to allow those systems to perform queries (such as Z39.50 or SRU/SRW) and to harvest data for their own uses.
- Permalinks Addressable URLs for catalog items that can be added to bibliographies, reading lists, etc.
Requirements (raw)
In this area you can dump any raw requirements, cut-and-paste from email or documents, or just toss off an idea. As raw requirements are moved up to the edited Requirements area, they should be deleted from here. The items here should be seen as work items for folks editing the Requirements.
May, 2007 kc
The list below is from the talk given by Prof. Tim Burke at the first public meeting of the LoC Future of Bibliographic Data group.
- tools that recognize existing clusters of knowledge; if you find a book using lcsh, you probably already know it existed. tool that recognizes the conversation the book was in. those that were written after the book came out and have continued the conversation.
- tools that know lines of descent; chronology of publications; later readers determine connection between texts
- tools that find unknown connections (full text search; topic maps?)
- tools that produce serendipity -- hidden connections.
- tools that inform me of authority
- tools that know about real world usage (those who bought x bought y; how many people checked this out?)
- tools that know about the sociology of knowledge; the pedigrees of authors: who were they trained by, how long ago; how trustworthy is this institution?
| Dec. 12, kc added Amy Ostrom's long list. There will be duplication here, so some editing needs to be done and more items moved to the "cooked" area. |
- I want it all in one place with option to see more or less (if it's on a booklist let me know, if it has reviews let me see them, put it all in the same place; if the library has it in audio and book format, put it in the same record!, seriously, if one type isn't in, I'll take another format - I don't want to click on 15 records just to find something; I also want to be able to hide some stuff if it is too cluttered)
- I want descriptions, dangit! And why does no software exist for integrating series information in the catalog?? I want to know what the next book is! (nothing like clicking on something you think you know what it is, and then it really isn't. I HATE that for something I might be interested in, I have to go to Amazon first to find anything)
- I want to see related/similar materials (I want a smarter version of http://www.literature-map.com/, either to graphically display the closeness of the book/author, or to at least list what others think are close)
- I want to make wishlists and my own booklists (heck, if I read an awesome series, I want to let others see these books if they share similar tastes; also I may not have time to read right now, but doesn't mean I want to forget a book I found that might be worth reading later)
- I want pictures! (I am visual, I'll know it's the right book if I can see it first; I want to see a sample of of the content as well, but would settle for a description)
- I want suggested searches and ways to narrow or broaden the search I made (if I can't remember the name or misspell it, I want it to act like Amazon and pull up suggested spellings or related searches, also broken down by category)
- I want the search to pull up the RIGHT materials (rank by popularity would work better than what item was last cataloged; Amazon is very good with its algorithm, it's not that hard to replicate - we can record how many times a record was viewed and how many times it was checked out, we know its publication date, we know its format, why can't we organize the search better?)
- I want an RSS feed for new items based on a search query (heck yeah I want to know what just came in without going to the catalog every day so I can get my hold on it ASAP, but I don't want to know EVERY item that is purchased, only what I am looking for; great for current awareness as well)
- I want to see the newly available items, especially in DVDs, CDs, and games (not just an RSS feed out, but actually on the site!)
- I want to know how long the wait list is, in days and or queue location (if it's too long, I'll just go buy the item)
- I want
permalinks, so I can link to a book from my blog to the catalog instead of to amazon or remember easily how to get back to it without running the search again (I am all about promoting the library, but Amazon is better than the library could ever be with marketing and promoting, let's take their example!) -
- It would be awesome to create my custom display, so I see what I want in the color I like (okay, it's a stretch, but it's all about customizing and personalizing these days)
- I want a map to show me the general shelf I might find my item (so many times an item was pulled out of the general collection and I pull my hair out in frustration)
- I want a library where I only have to sign in once, ONCE! (in my library catalog, every time I place a hold I have to enter my information; I log on, and I get signed out after maybe ten minutes of idleness - I'm probably surfing Amazon to find the RIGHT book...)
- I want to be able to turn on alerts for things like service outages, due dates, and overdues with quick access to renew, let alone modify my account profile and add password hints... (people are very forgetful)
So, from these wants, here is a basic (non-comprehensive) list of features we need to build a better catalog:
1. XML format
2. More (and better) content
3. More pictures
4. Smarter search engine
5. RSS on the fly
6. Commenting!! Commenting!!
7. User accounts
8. Single sign-in
9. User created lists/content
10. Permalinks
11. FRBR 2.0
12. Highly customizable interface
13. Highly user-friendly account settings/options
14. Smart spell-check aka related spelling/search terms
15. Organizable search results
- Be able to share our data with others (e.g. Google, any web services, patrons...). Allow users to create their own interfaces to the underlying data.
- ability to extend our systems beyond MARC. It's simply remarkable how we've allowed our vendors to lock up our systems, giving us only what they can develop and package up for us to buy.
- What do users want? Has anyone made a study? Do vendors know?
- Who are the stakeholders? 'Users', obviously, and cataloguers, but who else? System vendors? System owners? (whether these be universities, city administrations, or whoever) Others? What about businesses like online bookstores? Publishers? Other businesses? What about future users? (any system we design will either have enough flexibility built in to evolve with user needs, or it will be replaced by some other system)
- The OPAC has to show you completely current availability information from the circulation component---checked out, checked in, in transit, how many holds, etc.
- It needs to contain things that go beyond the things owned or licensed by a library. For example, in an academic library it might very well contain articles from scholarly journals or items harvested from the 'Net. (If this is true, then the term "catalog" may be mis-leading.)
- It needs to include services beyond find and identify. Some of these services might include review, annotate, download, manage citation, tag, facilitate social networking, etc. In this way the "catalog" becomes more of a tool as much as a searchable list.
- Support for multiple search interfaces, allowing libraries to enable patrons with greater or lesser skill to search the catalog in greater or lesser depth, rather than requiring libraries to struggle to make a single interface that would serve the needs of all patrons.
- In order to meet user's expectations, at least one of the searchable interfaces needs to be as simple as Google's interface. One box. One button. The old CARL character-mode interface offered users a choice of "words" or "names" - that's it. Enter words, get bibs and holdings. Enter names, get a list of names. Pick a name, get a list of bibs and holdings. Rather than have a flexible user interface, CARL stripped away detail to the point where you didn't need complexity or flexibility. What they appeared to want was a catalog anybody could use without explanation.
- Faceted navigation, spell-check, grouping functionality (e.g. FRBR).
- A documented standard for statistical data such as holdings and circulation transactions. Ideally also a translation guide between that standard and the measurement methodologies used by major ILS vendors, so that current numbers can be compared to historical numbers.
- Back-end function: The ability to automatically gather statistics about workflow, as statistics are now automatically gathered for circulation.
- Be written in code that is not hardware-specific. If the catalog application can be ported among several hardware platforms, its chances of catching on and surviving will increase dramatically.
- If libraries want to build their own next-generation catalog rather than waiting for vendors to do it, the project might need to be a joint effort among several major libraries, because multiple sponsors could provide more resources and keep the new catalog from created only for the needs of one specific library. The advances in communication technology that have made the open source community possible can help here as well.
I want several views, based on availability:
Some searches are about what I can find RIGHT NOW. Others are "does this exist anywhere"?
For an OPAC, do we really want one catalog per library? I belong to 3 libraries, 2 of which are consortia...
- Currently checked-in at this library
- Currently checked-in at my favorite libraries
- Available (due back or borrowable including via another library) within n days
- Everything in the bibliographic universe
I need to be able to select availability restrictions BEFORE or AFTER a search. My availability option should persist for my session or until I choose another availability view.
- Highlight (red, bold?) search terms in the results so I know WHY this record was returned to me (not some word in a 505 note somewhere that I can't see on the display: and I don't know why my search has returned this)
- More detail for the Spell check/Stemming area: SOUNDEX/METAPHONE (whatever), particularly for authors; variant spellings for British/American and other words (transparently!) labor/labour color/colour(particularly important for English-Canadian libraries) 19th/nineteenth etc.