[WEB4LIB] link annotations

[Robert Tiess]

"Martin, Julie" wrote:
> We've got some debate going on about the value of
> providing annotations for the resources we link to
> on our library web site. We link to a large number
> of databases, newsletters, electronic journals,
> and web sites relevant to our company.

Dear Julie,
Annotations are extremely valuable.  As someone who has created many web guides over the years, privately and professionally for either staff members or public use, I have several things to say on this subject, much of which you likely considered to some degree with your colleagues.  I know from experience many people share the questions you raise, so I'll go on a bit here in hopes of possibly helping you and anyone else considering these important issues.

Several things govern how far annotations can go, including your target audience, the number and types of resources linked to, the intelligence and intuitiveness of your web site/web guide's structure, and current (and future) available staff resources that may be available for database maintenance purposes (covering both data entry and description/URL updates) - when annotating, you must always think ahead to when you will have to update descriptions.  Early planning can reduce maintenance time and make your guide of even greater use, since it will be less prone to obsolete information, although nothing will eliminate the concern of broken links: documents will always be on the move, and information will always be subject to modification or removal.

Many sites, like Yahoo.com or Hotmail.com, don't require substantial annotation.  Each annotation should be as descriptive as necessary to give the user a brief but fair understanding of what is available at another site.  This saves the user time and provides better direction.  If such resources are primarily for staff, it will save the company staff time and money (less surfing = getting more work done).  If the links are for public use, further annotations may be necessary to bridge the gap between staff knowledge and the level of public awareness one cannot safely assume (e.g. many users would want to know what is available at a site called "Hotmail" before clicking that link).

The structure of your web site/guide also affects how descriptive annotations must be.  For example, if you have a section called "Free E-mail Providers" and stick "Hotmail" under there, no further annotation may be necessary.  A better schema can help eliminate the need for further explanations.  Breaking general categories, such as Databases, down into subgroups will helpful, such as "Business Databases" and "Health Databases."  Context simplifies things greatly.

Annotations can go too far.  Directional annotations to sites known not to change often are okay, such as telling a user to perform certain login tasks (enter user name/password) or click on a certain link/button that can't be linked to directly possibly due to a cookie-based session time-out or a remote CGI/PERL handler that acts only if a referrer URL originates within that site (usually to ensure a user follows a predetermined path at that site).  Beyond that, directional annotations may require much editing later on, after sites evolved (or devolved, as might be the case).

Summary annotations listing what can be found at a site are very usually helpful - as long as one doesn't get too specific.  Describing a site often turns into a balancing act between specificity and generality:  be too general and might as well not annotate it at all.  Be too specific and that's all the more you have to perform regular human reviews of hyperlinked resources to ensure what you described is still true.  Link verification in that sense can't be automated.  Annotations beyond one or two sentences get problematic when you link to many sites.  For practical purposes, annotations may be proportional to the resource guide (smaller guide, more descriptive annotations and vice versa).

Some other things to watch for in annotations:  Try not to editorialize.  Some of the more popular web directories engage in this, and when annotations turn into reviews, users are no longer making decisions based on objective observations.  Sometimes it's helpful to warn a site is "graphics intensive" (slow loading for users with dial-up connections) or uses something that might prevent certain users from accessing it (such as a plug-in).  If a site is that bad, you probably shouldn't link to it.  Including "hard coded" date ranges, such as a database covering a period of 1987 - 2001, will likely a year from now require revision; relative date ranges require less maintenance (e.g. "late 1980's to present").  Be cautious when describing certain services as "free": sites may have offered something free at the time you wrote the annotation, but we know how quickly things change online: consider how many "Britannica.com - free encyclopedia" links will need to be updated.  No on!
e wants to misinform users.  Also avoid site-to-site comparisons such as "like Yahoo"; site content and structure may change at any time.

Another way to look at this is to examine the extremes:  plain link lists versus utterly comprehensive site descriptions.  The first is of little or no value to the public (but of value to staff members who just need quick access to frequently used sites), while the latter is useless because it gives too much information for a user to make timely decisions.  Annotating as much as each site calls for may be the best approach.  I have also designed guides containing levels of description, so the user can click a "read more" link if necessary.  While annotations can be placed on separate pages, doing this places important information one more click away.  This is a more ideal technique for staff members, who would need to read about a site only once or twice before knowing what they need.  Developers of public sites should not assume users will know (or want to go) the extra click to learn otherwise critical information: elsewhere, consider how many users actually bother to cli!
ck on the About link of a site to determine the author(s), possible agenda, bias, currency....  Your annotation will help someone choose whether or not to incorporate information from an online resource into their research.

As you see, there is no one single perfect approach for annotating.  It's a mostly discretionary act, but more importantly it calls upon your training as an information specialist, to carefully consider and competently describe resources as they couldn't be by commercial sources with either financial motivations or staff not schooled in acquisition, classification, critical thinking, resource evaluation, and weeding.  This is why I continue to believe libraries can and should take over where search engines and commercial directories fail: to create the structure, logic, clarity, and intuitiveness few untrained persons and no web crawler or search engine will provide.

Compare Librarian's Index to the Internet (http://lii.org), one of the best annotated and most intelligent general web guides ever established, to any of the more popular directories, and the differences should astound anyone, especially when you realize LII is a non-profit project yet of substantially higher information value than the biggest corporate portals boasting countless sponsors and millions of dollars in revenue.  And that's what it's all about in the end: information value, information you can use now, information from sources you can trust, *informative* information that saves people time, energy, and maybe even money.

If you have any comments or questions, please feel free to contact me privately at my e-mail address below at any time.  Best of luck with your project!

Robert Tiess
rjtiess at warwick.net