Sarah wrote:
I don't mean that we should only use sources that are *easily* available, because that would be very limiting, but they must be reasonably available in the English-speaking world. I can't be precise about what I mean by "reasonable," but broadly speaking the man in the street should not have to move mountains, or pay a great deal of money, to obtain the material if he's determined to find it, and the source must be a published, citable one, not something circulated within a limited group of people.
Even there one must be flexible. If one of our objectives is to free this kind of information we are being pro-active in using it as references. If it is not reasonably available to the man on the street we should be looking for ways to make it available rather than suppressing references to it.
Even when we refer to "something circulated within a limited group of people" we can't be too rigid. I have, for example, a large bound typescript record of the papers presented at an international conference held in Utrecht in 1950. I suspect that its circulation was limited to conference attendees. I don't know if any of the papers were published elsewhere, or how many other copies survive. It's still recent enough for most of the papers to be protected by copyright, but old enough for some of them to be orphans. Does it advance human knowledge to suppress the material? Isn't it better to find ways to make it available?
The other issue is that of Original Research, and where "collation and organization" becomes OR.
The issues of availability of sources and OR merge when a Wikipedian explores a database that isn't widely available, as Zero was suggesting, adds his own interpretation of his research on that database to an article, and then calls the database his "source," even though there isn't a single document in that database that actually supports what he's saying. In other words, he has no citable, published source. That's a violation of NOR and V; it's a misuse of the primary source (the database); and it's arguably not reasonable to expect the man in the street to find a way to duplicate the research.
But otherwise, yes, I agree it's best to keep the two issues -- OR and availability of source material -- separate.
If you really believe they are separate issues then don't try to run them beck together. I like to look at databases as some kind of metastructure of sources. There is nothing "primary" to such databases at all. Whether Zero is correct to call them his source depends on how strictly you want to define that word.
Ec