Sure, and what I'd really like to see is a critique of functionality --
for example, what does it mean/what do you get when you click on an
author's name? What might you expect to get? And what functions are
missing, and why? (e.g. because not provided in library metadata, or not
interpreted by systems....) I think this is will be a difficult analysis
because there are interface, storage and retrieval, and metadata
problems that are deeply intertwined. I want to get down to an analysis
of where the metadata doesn't provide what is needed for the particular
functionality -- because that's my own interest, not because it's the
only important thing.
I'd like to include LibraryThing and Open Library, both of which
function in many ways like a catalog, but may have made different
assumptions being outside of libraries. However, they have relied on
library data to some extent, and therefore will suffer some of the
data-related failures.
kc
Stephens, Owen wrote:
> I wonder if rather than getting Tim to critique each different system
> - much as this would be very helpful ;) - we could agree a tag and
> interested parties coils blog and tag the results?
>
> How about #catcrit?
>
> On 15 Mar 2009, at 03:09, "Ross Singer" <rossfsinger_at_GMAIL.COM> wrote:
>
>
>> Tim, I think this is fantastic way of exposing the problems with
>> current library systems. Especially the way they generally display
>> the MARC verbatim, basically.
>>
>> That being said, critiquing Horizon is sort of like me breaking down a
>> video of me in a boxing ring with my two and a half year old. It's a
>> terribly one-sided fight.
>>
>> We all know that the last generation OPACs suck out loud. It's why
>> this list exists.
>>
>> I would be much more interested in seeing this treatment applied
>> towards Primo, VuFind, AquaBrowser, Encore, etc. The supposed "next
>> gen OPACs". After all, they're the systems we're all hanging our hats
>> on, scrambling to migrate to, etc. I realize that you chose BPL for
>> specific reasons, although you could probably take a stab at UMich's
>> VuFind implementation to help with the familiarity.
>>
>> Here's my guess: design and "web" wise, these will fare considerably
>> better (sessions in the URLs, general design, etc.) than your Horizon
>> example.
>>
>> However, I'll also wager they'll all fail majorly when you begin to
>> dissect how the records are actually displayed/used/etc. Interaction
>> between records. Formatting of fields. There is probably very little
>> departure, in reality, in the net result as long as they're still
>> basically regurgitating MARC.
>>
>> Diane is absolutely right, we're really just putting a nicer finish on
>> the turd until we stop thinking about "records" and "display intended
>> to print on cards".
>>
>> -Ross.
>>
>> On Sat, Mar 14, 2009 at 4:27 PM, Tim Spalding <tim_at_librarything.com>
>> wrote:
>>
>>> It seems to me that we go round and round on various issues. A lot of
>>> it resembles a "religious debate"--a term of art in software and
>>> design.
>>>
>>> One way to break religious debates is to talk about something
>>> concrete. So, I'm going to plop something on the collective "table,"
>>> offer a brief critique. Won't you join me?
>>>
>>> I've chosen the "detail" page of a book, the first "Obama" book that
>>> came up in the BPL[1]. I propose to critique it as follows:
>>>
>>> *I'm going to critique the page, not the whole system; I want to keep
>>> things focused. As such, I'm not going to critique the top part of
>>> the
>>> page, but just the part "below the chrome."
>>> *I'm going to list everything I think is wrong with the page, and
>>> offer brief commentart on it.
>>> *They use the HIP OPAC, one of the most common, but not the worst.
>>> *I punch because I love. I picked BPL because it is my favorite
>>> public
>>> library. I love them to pieces, and indeed I think their catalog is
>>> better than many.
>>> *This email is very long--10 typewritten pages. That's because the
>>> catalog has a lot of problems!
>>>
>>> Here's the page: http://www.librarything.com/pics/blog/ngc4lib2.png
>>>
>>> ### LEFT SIDE
>>>
>>> My criticisms:
>>>
>>> 1. The page is session-based. That means we have to discuss it by
>>> screenshot, it can't be spidered, it can't be bookmarked, it can't be
>>> sent to a friend and so forth. In my opinion, this is a catastrophe
>>> for libraries.[2]
>>>
>>> 2. No permalink. Despite #1, there are tricks to link to many (but
>>> not
>>> most) HIP pages, and other tricks that can link to all. The page
>>> COULD
>>> include a permalink, with icon--a familiar feature of sites like
>>> Google Maps. It does not.
>>>
>>> 3. The "Holdings" section on the left is misleading. The divet and
>>> other structure implies that the things underneath it are
>>> subheadings.
>>> I'm frankly uncertain if that's the intent. Maybe the list is of
>>> "holdings" related to the item. If so, "Holdings" is a very weird
>>> word. It's a weird word even in a librarian context--a link to a
>>> review is not a "holding." And it's weirder in a non-librarian
>>> context, where "holdings" doesn't mean anything at all.
>>>
>>> On further investigation I see that "Holdings" referred to the fact
>>> that, along with all the other info, the page I was on had the
>>> holdings info. The design is confusing.
>>>
>>> 4. The "Fiction and Biography" functions strangely. First, it's
>>> meaningless--what am I to expect that it does?
>>>
>>> My first guess would be that it would take me to other examples of
>>> Fiction and Biography. It does not. Instead, it takes me to a list of
>>> headings, "Genre" and "Topics." Both have subheadings, like
>>> "NonFiction" and "Politician." On Safari the headings--although
>>> black--turn into links when you roll over them. On Firefox they
>>> don't.
>>> In neither do the links go anywhere.
>>>
>>> The problem is probably technical. Whatever.
>>>
>>> 5. "Library Journal Review" works. Mostly. I searched for the review
>>> elsewhere, and Barnes and Noble has it. Barnes and Noble preserves
>>> the
>>> paragraph structure of the original, as well as italics and other
>>> formatting. The BPL has it as one giant text lump.
>>>
>>> 6. "Summary" is somewhat confusing insofar as there is a "Summary"
>>> field on the right, in the "book information area." The two are
>>> different. The summary works, but it's also somewhat "undigested." It
>>> ends with the non-sentence "a website where updates and comments may
>>> be posted as the campaign progresses: http://obamapolitics.com Book
>>> jacket."
>>>
>>> Why do libraries, which, if it stands for anything here, stands for
>>> sophistication and exactitude of metadata, allow thesse unformatted,
>>> half-gramattical text-blobs.
>>>
>>> 7. "Table of Contents" works pretty well. It's a bit odd, though,
>>> insofar as the same content is presented in the "Contents" field of
>>> the book-info-field.
>>>
>>> It took me a while to untagle the relationship between the two
>>> fields,
>>> though. It's non-obvious. Eventually I figured out that certain data
>>> (eg., the introduction, page numbers) were stripped out and returns
>>> replaced with "--". (See later for my rant against that idiotic
>>> typographical device.)
>>>
>>> 8. "More by this author" works as you'd think, but there are still
>>> problems:
>>>
>>> *Clicking on it takes you to other works by the author. In this case
>>> there are some. In many other cases, there aren't. When there isn't,
>>> it takes you to the record you are on! (Most users will record that
>>> as
>>> "I clicked on the link and nothing happened. So I did it again.
>>> Nothing. So I left and went to Amazon where the website works.)
>>> *The Last-First format is a fossil of the "dictionary catalog." In
>>> other book contexts--book covers, spines, bookstore displays, Amazon,
>>> B&N, LibraryThing, publisher websites, etc.--authors are First-Last.
>>> Only bibliographies still use last-first, precisely because
>>> bibliographies require "dictionary order." No such order is needed
>>> here. It looks fussy.
>>> *It's unclear to me why the author's first name needs to be followed
>>> by "1969-." Not even bibliographies do that. It's probable that
>>> libraries are in the pocket of the gravestone industry.
>>>
>>> 9. Subjects. Some problems:
>>>
>>> *Subjects are in Last-First format (eg., Obama, Barack). This is
>>> unusual--elsewhere I see his name as First-Last. There must be a good
>>> reason. Surely it is because it's an alphabetical list. Whoops, it's
>>> not. There is no reason for it.
>>>
>>> *The links *look* hierarchical, but they aren't. Given "United States
>>> -- Race relations -- Political aspects" you'd think you could click
>>> on
>>> any step of the hierarchy. You'd think wrong.
>>>
>>> *The links take you to a dictionary list of subjects, including the
>>> one you clicked on. You have to click it again to get something. That
>>> is, the link doesn't take you where you want to go, it takes you to a
>>> list of thinks, including a link to where you want to go. Did humans
>>> design this?
>>>
>>> *The use of "--" to indicate hierarchy is non-standard. The rest of
>>> the information work uses ">." It's unclear why libraries think the
>>> most basic web conventions must be ignored.
>>>
>>> *Whoever decided on using "--" in a web-product should spent five
>>> minutes with the Chicago Manual or Words into Type. "--" is what
>>> typewriters used for the em-dash. It didn't exist before typewriters,
>>> and it has no reason to exist now, when every computer and most
>>> cellphones are capable of the em-dash.
>>>
>>> *Since we're being persnickety, it's unclear why subjects end in
>>> periods. They aren't sentences. Punctuation, like "--" and "." have
>>> meanings. Misuse isn't a big deal, but it decreases confidence and
>>> tires the eye.
>>>
>>> *2001- should be an en-dash. Okay, I'll stop.
>>>
>>> 10. "Browse catalog by name" works okay, but it's unclear why it gets
>>> only one entry, and that's the author--who already got a link. Given
>>> the term "browse" I'd think that the link would put me in the middle
>>> of a millions-long list of books sorted by author. No. It does the
>>> same thing as the "More by this author" link.
>>>
>>> 11. "MARC Display" is weird. But at least it's small. Now, onto the
>>> main part of the page!
>>>
>>> ### MAIN PART
>>>
>>> 12. "Barack Obama : this improbable quest / John K. Wilson." Is weird
>>> in at least four ways. Together they reinforce the impression that
>>> the
>>> library catalog is arcane and fiddly.
>>>
>>> *The string "Title [slash] Author" is a library convention. In other
>>> situations, title and author are distinguished either typographically
>>> (as on a cover) or with the words "by."
>>>
>>> *In real life (except in France), book titles employ capital letters.
>>> As I've said before on this list, when LibraryThing started showing
>>> library titles, users complained that the site was "broken."
>>> Something
>>> was causing book titles to lose their capitals. Funny? Alas, the
>>> jokes
>>> on libraries.
>>>
>>> *In real life, colons don't have spaces before them.
>>>
>>> *In real life, author-title lists don't end in periods.
>>>
>>> 13. All text from this point goes underlined when rolled over. But
>>> it's not clickable. This makes no sense at all. If LibraryThing did
>>> this, I'd have ten bug-reports inside of a minute. I wonder if people
>>> report this, or if the general atmosphere of brokeness prevents it.
>>>
>>> 14. "Publisher," "Boulder, CO : Paradigm Publishers, c2008." What a
>>> peculiar string. The publisher is "Paradigm Publishers" but their
>>> location is listed first? And why call it "Publisher" when it
>>> includes
>>> publisher-town, publisher-state, publisher name, copyright symbol (in
>>> case you thought it might be public domain?), and publication year.
>>>
>>> Order implies importance. On what planet is the publisher's location
>>> the second-most important fact about this book?
>>>
>>> 15. "ISBN: 1594514763." Is this really the second-most important fact
>>> about the book? To whom? I know it's shocking, but most readers don't
>>> know what an ISBN is. The rest don't care.
>>>
>>> And ISBN *might* be a useful way for a knowledgeable user to jump
>>> from
>>> Amazon to a libray catalog. But they'd have to get the right edition.
>>> The rest of the time, the ISBN is trivia for stockboys.
>>>
>>> 16. "Description: vi, 210 p. : ill. ; 24 cm." As others have said,
>>> this is a meaningless jumble. It doesn't merit the title
>>> "description." It's junk.
>>>
>>> The patron *might* want to know how long a book is--so "210 pages"
>>> migtht be useful. I'd even be fine with "216 pages."
>>>
>>> The patron might also want to know that the book had photographs. It
>>> would be better to know how many, or even to get a list of them.
>>>
>>> 24cm is wrong in about ten ways. First, although "the most European
>>> city in America," Boston is still part of the USA. In the USA we use
>>> inches, not centimeters. There's there's the issue of one
>>> measurement.
>>> Is can't be width. Is it height? Width? Maybe it's like TV and
>>> computer monitors. That must be it.
>>>
>>> 17. "Target audience: Adult." This is useful here. There are a lot of
>>> kids books about Obama. I'm glad this isn't one of them.
>>>
>>> 18. "Summary." Fine, except for point six, above.
>>>
>>> 19. "Contents." Fine, except for point seven, above.
>>>
>>> 20. "Bibliography: Includes bibliographical references and index."
>>> This is marginally useful. For "bibliographical references," I'd use
>>> the plain-jane word "bibliography."
>>>
>>> It's funny that this is spelled out in prose but "ill." isn't.
>>> There's
>>> probably some good reason.
>>>
>>> 21. The "Copy/Holdings information" box has some problems:
>>>
>>> *I'd love to be able to click on locations to find out where they
>>> are.
>>> Although a long-time Boston resident, "O'Bryant School of Mathematics
>>> and Science" means nothing to me. I guess I'll have to ask at the
>>> desk.
>>>
>>> *The "Collection: Nonfiction" confuses me. I wasn't aware libraries
>>> were divided that way. In fact, they aren't. But there probably is a
>>> fiction section, that includes *most* of the fiction.
>>>
>>> *Capitalization is almost random. "14 day loan" but "In Library"?
>>> Why?
>>> Or take the capitalization of the section header, "Copy/Holdings
>>> information." The other place with a similarly-styled heading uses
>>> title case, "Related Information." Small inconsistencies make a site
>>> look sloppy.
>>>
>>> 22. The form below is literally backward. It's formatted like this:
>>>
>>> Format: ( ) HTML ( ) Plain text ( ) Delimited
>>> Subject: __________________
>>> Email to: __________________ [SEND]
>>>
>>> When I reach the third line, I gather it's an emailing form. Why
>>> doesn't it look like almost all other email forms on the web--the
>>> ones
>>> that START WITH THE EMAIL ADDRESS? Also:
>>>
>>> *Whoever designed this form didn't look at how you email things on
>>> any
>>> other site! Do I need the subject field? If I do, why can't I write a
>>> note.
>>>
>>> *What is "Delimited" anyway? I have no idea.
>>>
>>> *The form cuts off the title, into "Barack Obama : this improbable
>>> quest / J". The form field allows only 40 characters, but is visually
>>> larger. This is confusing and completely opposite how most web forms
>>> work. Anyway, why are they only allowing 40 characters--bandwidth
>>> costs?
>>>
>>> 23. "Next Reads." I gather Next Reads is much liked.[3] But this
>>> "advertisment" feels intrusive and over-prominent. It certainly
>>> doesn't fit in with the design at all. The line "Sign up for email
>>> book suggestions in your favorite genre!" *may* relate to the icon
>>> here, or it may not.
>>>
>>> 24. "Did you know? Many items held by the BPL are not listed in this
>>> catalog. Find out about all of our catalogs."
>>>
>>> *This notice is not visually separated from the line above, about
>>> NextReads. Are they all part of the same notice. All separate?
>>>
>>> *The notice is certainly unfortunate. If all their stuff isn't in the
>>> catalog, they need something like this. But it certainly raises
>>> doubts.
>>>
>>> ### Final Points
>>>
>>> 25. The design is unappealing and slapdash. Some examples:
>>>
>>> *The information architecture of the left-hand side is all weird.
>>> I've
>>> mentioned the divet and the "--"s in the subject. But what about the
>>> stray horizontal line in between "Table of Contents" and "More by
>>> this
>>> author"? Is it necessary? Is it attractive? Did somebody's teenager
>>> design this?
>>>
>>> *The "Add to My List" and "Hold this for me" buttons, although on the
>>> far right, are somehow creating extra space between the book title
>>> and
>>> its information. To an untrained user it's just another tiny mark of
>>> inferior quality. To the trained web developer it's evidence that
>>> someone doesn't understand floats.
>>>
>>> 25. Font sizes
>>>
>>> *The most important information--the book info and the holdings
>>> info--are in the smallest fonts. That's crazy.
>>>
>>> *Apart from that, font sizes and styles are slapdash. The title of
>>> the
>>> book is less prominent than "Related Information."
>>>
>>> 26. Accessibility
>>>
>>> *The page fails all levels of all accessibility tests. Five years
>>> ago,
>>> when I made school software, I paid close attention accessibility.
>>> Governments all require it. How did libraries get to opt out?
>>>
>>> *Test aside, nobody has looked at basic accessibility issues--
>>> semantic
>>> coding, order of information, tab-order, alt-text, etc.
>>>
>>> 27. The great bullet problem
>>>
>>> Finally, as web developer I have to mention one thing that, when I
>>> found it, made me laugh out loud--and I don't usually do that.
>>>
>>> The bullets on the left--the giant, ugly bullets that don't quite
>>> align right--are not an unordered list (<ul><li>...</li></ul>, etc.).
>>> They are instead a table, with two columns--someone's attempt to
>>> produce a bulleted list, without using the HTML markup for... a
>>> bulleted list!
>>>
>>> To get it, the list was "tableized." The left-hand column is for the
>>> bullets. But instead of printing the unicode for a bullet, using a
>>> graphic or wahtever, the left-hand column is comprised of
>>> single-entry, no-content unordered lists. Apparently someone at
>>> SirsiDynix thought that <li> </li> was a trick to get a bullet.
>>>
>>> [1] The BPL is having some sort of deep problem. Most of my searches
>>> turn up page after page of blank records. This was the first non-
>>> blank
>>> one.
>>> [2] Not being "on the web" is, in my opinion, the single most
>>> important factor that drags libraries down in the internet age, and
>>> therefore a great threat to library success, library jobs and indeed
>>> to education and democracy. But hey, what do I know? Maybe the rest
>>> of
>>> the web is wrong and libraries are right!
>>> [3] I don't know the product very well, but I am a fan of Novelist
>>> and
>>> its people.
>>>
>>>
>
>
>
--
-----------------------------------
Karen Coyle / Digital Library Consultant
kcoyle@kcoyle.net http://www.kcoyle.net
ph.: 510-540-7596 skype: kcoylenet
fx.: 510-848-3913
mo.: 510-435-8234
------------------------------------
Received on Sun Mar 15 2009 - 14:13:07 EDT
