I am finding more items listed as sources in the FamilySearch.org Family Tree that are not sources at all. They are revealing as to where the people who entered the data obtained their information and that fact is helpful. But the place where you got the information is not necessarily a "source." What do I mean by this statement? The issue here functions on different levels. There are those who would support the idea that putting anything at all in as a source is a huge step forward from the past when no sources were listed, but I think it is important to move beyond that level and improve the quality of the sources we use to compile our unified family tree.
The first level, of course, is where users add or change the data in the Family Tree program without adding any information as to where the new information or changed information originated. The basic question here is: "Where did you get your information?" By answering this question, the user provides others with a way to evaluate the reliability of the change or addition and, where appropriate, examine the same document or other originator.
I had that experience just yesterday. I have an ancestor who is often referred to as "adopted." Personally, I had never found a source, i.e. original record or document, reflecting that characterization other than "rumor or family stories." Recently, I found another reference to the fact that this ancestor was "adopted," this time in writing with a reference to a church record. The record indicated that the researcher had referred to an "index card" that mentioned that the ancestor was "adopted." I could not find the index card, so I went back to the original record from which the index was likely derived. In this case, the original entry bore no reference at all to an "adoption." Now there is a mystery as to why the derived record had that notation, if it did. Presently, I have no idea how to find the "card index" referred to.
This illustrates a couple of important points about what we need to supply when we provide a source.
First of all and very importantly, the format of a source citation is not as important as the content. I don't really care what format was used to tell me about the card index example I used, but it would have been nice if the researcher had entered enough information to tell me where I could find the record she was referring to. The main idea behind providing sources is to give subsequent users or researchers the opportunity to find the record used. For example, telling me you got the information from your "personal knowledge" is nice, if there is some way given about how to contact you for verification of what was recorded. I fully realize that in some societies around the world, the only family history information available is in the form of oral traditions. If the source of this family history is listed as "oral history" from a certain family historian, at least I know where the information originated. The original informant may now be dead but that may be the only source for that particular information. Whether or not other sources exist for the same information is not the point. The issue is recording where you got the information.
Next, be as complete as possible. Don't assume that anyone knows about your source. In many cases, it is better to provide more information about the originating document or person and their location than less. I am not saying that "personal information" is not a valid source, but by giving me as much information as possible about how and where the information was obtained, I have, at least, a possibility of evaluating the information based on the reliability of the source. When a source is completely lacking, I must conclude that anything recorded is suspect and must be verified. If you copied the information from a Family Group Record or from an online family tree program, I can determine whether or not I think the information needs further documentation and research.
Lastly, provide a copy of the original document whenever possible. This is more than a convenience. I can tell immediately, by looking at the document, if you copied the information correctly, decoded the handwriting or whatever. It does save time and indicates to me that you actually did some research and did not just copy the information.
The answer to the question posed in the title to this post is this: A source is not a source when it fails to give adequate information to evaluate the reliability of what is recorded as being derived from the source. Obviously, if no source is provided, the reliability of the recorded information is close to zero.