Invisible Cities logo

MARTIN LANGER, 2919-E N. Military Trail, West Palm Beach, FL 33409. Scholarly books - first editions - gay and lesbian literature. Write: martin at InvisibleCities dot com, or call him at: (561) 346-2002. 

This site consists of descriptions of products for sale, one product per page. At the time it was catalogued, each product was assigned an SKU number. These SKU numbers are presented sequentially in groups of 100 per page. Each SKU is a link to the corresponding product's individual web page. SKU's (and corresponding web pages) are deleted as products sell.

If the foregoing isn't very interesting to you, I'm not surprised. This home page and the pages to which it directly links contribute nothing to the utility of the site. They are here only because they are mandated by search engine guidelines as a condition for the content pages of the site being included in search engine indexes. In other words, from the point of view of a site user they are a distraction, yet without them users would not be able to find my content pages.

Granted, if I had money and inclination, I could publicize the site as a whole, hope that that would result in people discovering my content pages, and trust that once they became popular, they would be picked up by search engines. But I submit that's backwards. The relevance of my content to people's queries should cause it, via the proper operation of search engines, to attract people to my site. Advertising should not be needed.

As I see it, the problem here is that search engines which mandate hierarchical site structures -- homepages linking to categorized pages linking finally to content pages -- assume that for all sites, a hierarchical structure is a prerequsite to a quality site. I think this is wrong as a proposition and wrong-headed in it's preoccupation.

It is wrong as a proposition because it takes into account only one worthwhile site model: that of the collection. If a site is presenting information on a given topic, then comprehensiveness, balance and organization are desirable attributes. You would expect them in, say, an encyclopedia article or a museum exhibit. But might a list not be a worthwhile site model, too ? If, in a collection, things are all where they belong, then what if things are where they don't belong ? Let's call that the junkyard model. And, while it may be more prestigious to be a museum curator than a spare parts dealer, are not both functions important ? The curator maintains and augments the integrity of his collection. The junk dealer attempts to disperse unwanted parts to where they are actually needed. And since the principle governing the acquisition of spare parts is that items which are not where they belong tend to be undervalued, the last attributes one would expect of a junkyard are comprehensiveness, balance and organization. Instead, in a junk yard (or a junk yard website) one might seek variety and low prices.

And mention of junkyards brings me to the wrong-headedness of search engine's preoccupation with site quality to begin with. It is directories (like Yahoo and dmoz) whose business it is to evaluate overall site quality -- where that includes considerations like site architecture and attractiveness of layout -- not just because they've staked out the territory, but because human discrimination is called for. Search engines should concentrate on the rightness of particular bits of content as answers for particular queries -- in whatever unexpected context they may happen to be found. In other words, when things are where they belong on the web, one can trust to directories to navigate there in an orderly, hierarchical manner. It is precisely when things are not where they belong that we need search engines. Nor is it just ironic that supremely non-hierarchical search engines are mandating hierarchical organization for all sites they index. For 'junk yard' sites, it's actually illogical.

Suppose, for example, a bookstore gets a book on Babylonian dams and irrigation. If the store were asked to assemble an exhibit of books on Babylon, it would be appropriate to include it in a section on Babylonian agriculture. Given a focus, a defensible hierarchy can be constructed. But if the store were to shelve the book under History: Ancient Near East: Babylon, then if a customer looked for it under: Engineering: Civil: History, it would not be there. In other words, the categorization of an item according to one scheme, may deprive another category of the item. In online schemes, one may hope for multiple hierarchies to include the same item, but even the most persnickity search engine would not be requiring a single site to arrange its content according to all possible classification schemes. In other words, placing an item in a classification scheme may make it less likely to be found; so for a site which aims to disperse items which do not belong together in the first place, a presentation of each item on its own may be better than presenting it in a putative 'context'.

This is such a site. Having a home page and linked pages of groupings of items contributes nothing to the utility of the site for end users. Therefore, if you judge my site design, please pretend these search engine-required pages were absent but that my content pages had nevertheless been indexed by the search engines.

To summarize: when the content of a site, as here, consists of random products offered for sale, having in common only that they were acquired by the same merchant because they were previously underpriced, then attempts to impose an organization on the resulting hodgepodge may hinder efforts to find for each item independently the ideal buyer. Sites notable for their excellence as collections should be recommended as such by directories. Sites serving as spare parts dumps should have their constituents as widely disseminated as possible after having been picked apart by search engines. If you agree, I hope you'll join me in urging search engines to concentrate on locating content and to relinquish their too-narrow doctrines concerning desirable site architectures.

So, having now offered my apology, I dutifully offer up the mandated lists. Collective characterizations are not provided because the constituents do not (and could not, however arranged,) constitute coherent collections. Suffice it to say: they're books.

Items 1700-1799
Items 1800-1899
Items 1900-1999
Items 2000-2099
Items 2100-2199
Items 2200-2299
Items 2300-2399
Items 2400-2499
Items 2500-2599
Items 2600-2699
Items 2700-2799
Items 2800-2899
Items 2900-2999
Items 3000-3099



Invisible Cities logo