Singhals wrote: > It is tempting, oh-so tempting!, to list as a source for > great-grand's marriage something like > http://members.NASA.edu/web/web_page/this.html > and skip the wearisome step of VERIFYING the info in the real > records. > > The problem will arise (that's WILL, as in guaranteed) when, two, > six, eight months or a year from now when you try to go back to > http://members.NASA.edu/web/web_page/this.html > and get a 404- file not found, or a No DNS. Download the data. Note the address. Print it if you wish, and date the printout. But in any case, file an image. That way, if you self-publish or if anyone asks, you can simply assert that you got the data from the specified address on the stated date. And you will at least have a copy of the original. Verification is a concern primarily for outside publishers and researchers who desire evidence from sources as close to original as possible. Personally, I believe that it does little good to provide exact Web addresses for images online with plenty of offline accessibility as well (such as census data). They consume space, they can be found online or off by the same fixed identifier as conferred by the original distributor of the data, and as Cheryl points out they are subject to drastic change without notice. Under these circumstances, a standard offline reference with an addendum like, "image file "*.jpg", downloaded from Ancestry.com, 23 Apr 2003" may suffice. The online accessibility of census data is one condition with which citation standards have not yet caught up. [snip] > It seems to me that one of several other approaches could be (should > be?) taken -- cite the source (1) as "Jerry Murphy's website, Jul > 2001" or (2) as "Private Communication" or (3) [a personal favorite] > "I read this somewhere." (2) and (3) are unarguable, not open to > verification, and no worse than an invalid URL; (1) at least gives > you a fighting chance of determining whether the website you find > today is the one you saw last year. > > Cheryl <singhals@erols.com> The accepted form is author, "title," online <address>, date of last update, date of download or viewing. In truth, any author regardless of medium has the ability to make any of these elements difficult to recover (except "online," which is also unarguable); only the volatility of Web addresses is new. A related problem arises when data are distributed among several different documents, as with most "books online" and Web-rendered "databases" of the GED2* variety. You cannot instantly recover a data-containing page or image from such a collection without typing its address in full; if you cite different documents repeatedly, you'll wind up with a a bunch of links to the same root directory, which under most circumstances cannot be shortened and remain functional. You'd better use a word processor to organize your research results for public display, and avoid splitting related facts when you can. Austin W. Spencer "Austin W. Spencer" <AustinWSpencer@sdc.cox.net>