The purpose of the abstract is to provide a concise representation of the topic(s), research question(s), and main finding(s) of the FULL paper. And that is ALL it is. Writing a very good abstract is not easy, of course, especially if the maximum number of characters allowed is low. It is an art in itself to be such a good author of your abstract that you will attract many readers towards the full paper. But one must remember that an abstract is not an advertisement that should "sell" the paper maximally - exaggerations are not allowed. The abstract shall be written to the point, it should be informative, and it shall NOT exaggerate the content. And I'll say it again: it does not work as a stand-alone entity.

For such a purpose there is something called an "extended abstract", where results can be shown a bit more fully. That is a useful model for assembling short version of conference papers in the conference bag. But an plain abstract cannot replace a full paper and cannot be viewed as a stand-alone entity.

I agree with this & it has been a logical & shared concept amongst most of the colleagues I have worked with. 

** Is there any cited source that supports it? **

Agreed reading abstract alone might not understand the whole paper as different authors might focus / emphasize on different key points in their abstract.  Different journals / conference proceedings also restrict abstract's words differently from 150 to 500 words etc.  But I find Emerald publisher with its associated journals normally required author(s) to write their abstract in the following format / sections:

1) Purpose

2) Design / Methodology / Approach

3) Findings

4) Practical Implications

5) Originality / Value

6) Paper Type (e.g. research paper or opinion etc)

I find using this format can help readers / me understand a particular article as fast & as much as I can.

Not sure how to get the cited source to support this format but all the articles published by Emerald & its associated journals always come with the above format.  May be you can goto this link on Emerald & view some sample / free articles:

http://www.emeraldgrouppublishing.com/products/journals/journals_index.htm

Dear Kamal,

That is true in regard to developing the abstracts of oral presentations, and figures range between 20-70% depending on field. Oral presentations sometimes have the paper printed in conference proceedings, so you can get some depth of information if you have access to the proceedings. Also, they are presented to a fairly large proportion of the conference body, so you can assume that a decent amount of information has been provided to support the presented work.  Posters however attract little active attention, with only around 3-5 visitors cited as interacting with an individual poster at national/international events. In terms of volume, they vastly outnumber the amount of oral presentations but the information they offer is not available outside the poster session beyond the abstract. Moreover, they are less frequently developed into a published paper as they need considerably more work to go from the original abstract submission (@300 words), to a full article (@3-5000 words). Oral presentations tend to have more structured information & are more readily converted.  There is also a grey area in identifying what is a direct presentation development; i.e. a paper with the same title, presenting exactly the same data in text form & reaching the same conclusions. It is difficult to see how anyone has actually assessed these factors, but they are important if we are to claim that presentations are converted to 'published' material, & therefore the information is potentially available to the wider peer community.  More loosely, it is likely that later papers include coverage of the presented topic, but also give a wider perspective. They are also a more readily envisaged output & are probably intended anyhow, regardless if a topic (or part thereof) has been offered at a conference.  So, is such a paper a 'development' of the conference presentation, or is it a separate (but related) entity ....?

The question here however is:  Is there any cited source that supports the view that abstracts can be seen as only an indicator of the work undertaken,  but not as robust or reliable as a stand alone entity.

The purpose of the abstract is to provide a concise representation of the topic(s), research question(s), and main finding(s) of the FULL paper. And that is ALL it is. Writing a very good abstract is not easy, of course, especially if the maximum number of characters allowed is low. It is an art in itself to be such a good author of your abstract that you will attract many readers towards the full paper. But one must remember that an abstract is not an advertisement that should "sell" the paper maximally - exaggerations are not allowed. The abstract shall be written to the point, it should be informative, and it shall NOT exaggerate the content. And I'll say it again: it does not work as a stand-alone entity.

For such a purpose there is something called an "extended abstract", where results can be shown a bit more fully. That is a useful model for assembling short version of conference papers in the conference bag. But an plain abstract cannot replace a full paper and cannot be viewed as a stand-alone entity.

Original paper are more adequate than abstract. Publishing abstracts without extending them tp full paper is not good practice.  

Dear All,

An abstract is aimed only to give brief information (often without date and descriptive parts) on a paper whose message can be various regarding the subject and the deepness of investigation. Each of us should agree with Kamal: abstracts are mainly for orientation. They contain imperfect information compared to whole papers. Many abstracts are very (insufficiently) brief because editors try to economise charges. It is not correct to put side by side an abstract and a whole paper.

Dear @Nicholas, I do agree that abstract is only an indicator of the work! Furthermore, abstract is concise description of work on which basis potential readers comes and ask/pick a full paper. 

I do not have yet an example of such citation! Only full paper should be cited, not abstracts!

How to Write an Abstract seems to be very good reading/hints for good abstract and paper!

http://users.ece.cmu.edu/~koopman/essays/abstract.html

I would think it would be enough to write a couple conference organizers (preferably and ask the same question and then cite them (with their permission, and/or anonymized with something like "conference organizers of 6 major conferences agreed that abstracts are helpful for abc but are not an indicator of the actual paper content).

Interesting replies & one's which reaffirm a quite obvious position.  It is however interesting that the reality of practice shows us to produce abstracts as a sole output.  Why would this be, if we know they are inadequate as a standalone medium?