Comments on the SiteInspector checker

As of 2000-04-04, it seem that the SiteInspector service has been discontinued. The old address now points to bCentral, which has a service remotely resembling that of SiteInspector, but it can't even handle relative links when doing link checks! Oh well... I think I'll preserve this document as a history document.

SiteInspector is an online service for checking Web pages (HTML documents). This document briefly characterizes what it checks and how well it does that. Some instructions on interpreting SiteInspector reports are given here too. SiteInspector is a quite handy checker which lets you get analyses of your page from different points of view, but you need to know how to read the reports.

SiteInspector performs seven different checks and reports results from each of them separately. You can however turn off some of the checks by unchecking the checkboxes in SiteInspector's own user interface. (On the right you see a slightly modified version of the user interface, with the checkboxes sorted so that order matches the order of the reports you get.)

Next we shall briefly characterize each of the checks. (See also SiteInspector's own Quick Help.)

URL:
Browser Compatibility
Link Validator
Popularity
Load Time
Submit It! Readiness
Spelling
HTML Validity

Browser Compatibility

This analysis is useful to an expert, but it could easily be misleading. It flags some constructs which are not supported by some important browsers. It complains about the lang attribute and many other things which involve no compatibility problems. (Moreover, contrary to what SiteInspector says, Netscape supports lang no less than Internet Explorer; both ignore it.)

Link Validator

Checking that links works is an extremely useful feature. However, it seems that SiteInspector only says either "OK" or "Fail", and the latter could be just a temporary problem (e.g. host not reachable.) (Why can't it tell the HTTP response code? It surely must base its conclusions on them!) Moreover, it only checks at most 25 links at a time. But once you know this, and remember to ask for more checks, it's not a big problem, especially since SiteInspector is rather fast.

Popularity

The "popularity" estimate seems to be based only on links to the page in one data base, namely that of AltaVista. So it just tells you how many such "sites" link to your page which made their way into AltaVista. This tells something, but nothing very definite. And there seems to be the usual confusion between the concepts "page" and "site" here. (Apropos, looking at such information, i.e. pages linking to my page, need not be just an ego-trip. If someone is interested in my page, he quite possibly has links to other pages or pages of his own on the topic, perhaps something new and relevant to me.)

Load Time

This contains an estimate of loading times with some typical data transfer rates. Useful for getting a rough idea of how fast or slow your pages might be, but it seems that the estimate is based on a rather optimistic estimate, e.g. that 14.4K really means 14,400 bits per second as effective rate!

Submit It! Readiness

This report contains an analysis of the page's "search engine friendliness". Take a look at it, but the report really gives a misleading impression about the usefulness of keyword lists. Moreover, it generally overemphasizes the usefulness of meta elements - several important search engines pay no attention to them, so it is useful to take another look, e.g. with Delorie's Search Engine Simulator.

Spelling

SiteInspector seems to check the spelling of words according to (American) English orthography in a manner similar to several other services. Thus, it can be useful for catching typos. But it is a rather naive check, based on a list of words, and the list is somewhat wanting. The checker apparently does not recognize even plural forms of less common words although it knows them in singular! Additionally, it occasionally reads words wrong, e.g. the word "technically" as "chnically", so there are bogus entries in the report, too. This seems to happen for first words in parenthetic remarks.

Naturally, if the document is in another language than English, you should turn off this check.

HTML Validity

This does not mean checking the document using a real validator. The checking is however reasonably good, and the messages often refer to errors in the document in an understandable manner. Moreover, it gives warnings about features which are valid but not supported by all browsers (e.g. apostrophes around attribute values instead of quotation marks).

It also makes some erroneous claims, e.g. reporting the absence of <HTML> and </HTML> tags as an error.


SiteInspector can also be used via a simple user interface written by me. And there is a frameset version of the interface, too.


Date of last update: 1999-05-07

There is a Finnish version of this document - Tästä dokumentista on suomenkielinen versio: Kommentteja SiteInspector-tarkistimesta.

Jukka Korpela