We need to distinguish between a bad paper and a bad journal. Recent scientometric research has shown that the quality of the journal where one publishes matters less for citation impact than it did before the Internet era. So long as the journal is accessible online, if the paper is good, the citation count penalty today for having it published in a lesser journal is less than it would have been in the 1970s or 1980s. Of course, there will still be a price paid on matters other than citation counts. At many institutions, quality of journals remains an explicit criterion for research quality evaluation, whereas citation counts are often less attended to. It depends who is assessing you and where/ in what context. 

Another aspect to consider, is that many funding bodies have a moving window of publications they evaluate for a candidate's performance: for example, for an SSHRC grant, they focus on your publications in the last six years. That mechanism means that older sins eventually get forgotten. 

I have also heard of the practice of scholars' deleting from their CV publications they come to dislike. 

I attach a short paper of mine on the epistemic logic of the peer-review process - you may find it useful for thinking through these issues:

Article Error, quality, and applied geography: An editorial on process

This is a good question that I have thought about.  We are all not all in a position where our hard work, ideas, experience, expertise is supported by our company, agency, position or outside support.  At some point, each individual, researcher, technical writer who wants to share information or ideas for some reason, how best to do that.  Some have experiences and expertise as professionals, not as researchers per se, yet it is so difficult to share information and compete with those that have research connections such as universities, agencies.  As less well known such as students, small school researchers may not have an easy pathway to fund research or even if  done on personal time, a pathway toward publication.  Perhaps these online publishing serve a purpose, in that if you someday become more famous, your previous works may gain a higher return.  And sometimes work may be done, reports written with substantial data, and the information is lost as there is no effort to publish it or save it, and as time passes, the raw data, paper or even digital files are lost with time, moves, software changes, etc.  

In addition, some of us are not that great at either researching or writing, and perhaps these online publishing serve some purpose for that.  There is little to offer a step up in respect unless initially fortunate enough to study and work under a noted scientist, who is well published.  

I truly think the grant, peer reviewed and scientific community looses out of some expertise and content by supporting only those with clear research connections or developed funding support.  True, we may not all deserve full support for funding, but technical meetings, symposiums, recognized publishing avenues often require funding support to participate.  If you are a student, non-research positioned professional or a retiree as myself with the desire to contribute to and remain involved with the science in some way, the path to participate often requires funding from somewhere.  Some find it with cooperators, some perhaps might be approved by their company or agency, but the bottom line may be self funding.

Unless your position is in research or at a university, NGO, non-profit organization, seldom is there any opportunity for grants, sharing of tight research funds, special provisions or reduced or waived fees for those who want to publish, want to contribute to technical meetings or conferences, want to pass on experience, etc.  So perhaps Researchgate for me has become that temporary technical outlet, but online publishing may be another avenue where no funds are needed.  

Just like lawyers take some work pro-bono, why are not not more avenues for respectable publication outlets where they are wiling to waive or reduce fees?  The alternative I suppose is self funding, or if really good enough, writing a book.  

I think we all have written things in our career that certainly the wording is not up to our current standard, and in decades past, data had to be manually analyzed, plots drafted by pen and ink, hand typing, etc. that may look amateurish at best.  Unfortunately, some may have a few skeletons where work was not up to snuff, efforts did not work out, data was lost or had gaps, the project failed, or the subject was republished so many times to be embarrassing.  

I doubt that there are few researchers, former students or publishing professionals who would like a take back or redo.  And sometimes a redo is responsible as the original was either so bad or published to soon, before adequate data was collected.  The redo is a  good time to set it right, but how to destroy the skeleton of bad or incomplete work, I do not know how that might be accomplished in this day and age, with so much on line.  Anyone that has those worries, just forgive yourself, move on, don't repeat mistakes, apologize if necessary and only continue with quality work to the best of your ability.  

If abilities are average, you have desire to share, interesting data and case studies, technical reports, no funding support, no expectations to get revenues in the future from ideas, on-line publishing seems a reasonable approach to consider.  Perhaps Researchgate entries is another alternative, but how long it will be popular and maintained is unknown.  I do think that if researchers find significant information worth citing, whether it is a quality publication, on-line published, agency report, powerpoint presentation, personal communication or Researchgate entry, that researcher does owe the writer the courtesy of a citation, or if not allowed by the publisher, acknowledgement.

I think we may have concerns that on-line publishing is not necessarily peer reviewed, technically responsible or even mainly junk, publishers may be getting rich, the system is being overloaded and diluted, and information by some on-line publishers may require payment and not be free (as also true of some recognized publishers).  As long as these on-line publishers have full disclosure of how your information will be used and available to others, why not have this as an option?

For those researchers, publishers and content with very high standards or significant original works, royalties, fees, etc. for reading materials may be appropriate.  Those operating and functioning at that level deserve the ability to recover their costs and continuing payment for technically transferring their significant experience/expertise.  It would be nice to think like patents, with time, these might eventually be added to the information available to the public for no or minimal costs.  

It is probably appropriate for these on-line publishers to supervise their content for quality, and their degree of supervision will determine their reputation.  I would hope that they would remove any existing works desired by the author from their system with an agreement with the author that if the work is republished or content used for economic gains, the on-line publisher would get some cut.  Perhaps some wording along these lines can be negotiated as a condition of the on-line publishing.  If it was just poor work, no or only minimal costs to cover their time to actually remove from their data base might be in order.  With so much on the internet these days, who knows how difficult or costly a complete purge might be.  So if the author or on-line publisher knows it is very poor, irresponsible or even negligent work, ethically a self analysis or publishing should step in and not allow its entry.

I will be interested to see the other responses, but I think this is a topic of interest to any who have published, as the emails soliciting on-line publishing are increasing, and it would be interesting to hear from those who might have some horror stories to tell on how their work was used to their benefit or to their damage in some way.   

Thanks Dragos sir and William sir for your thoughtful response. However, I am unable to judge whether or not to publish graduate students primary research work in sub-standard quality journals? I have elaborated about the quality of the journals for which i am using the term substandard. Though my students research works are based on vary small area but still much better than the research papers that are published in so called online free paid ISBN journals (about which I have written in the question). These journals at this stage do not even follow the ethical publication procedure, say for example plagiarism check, checking of mathematical/statistical formulas and function, symbol verification, language editing, cross-cheeking of references etc forget about authentic peer review. If those articles be diverted to mainstream journals say with impact factor less than 0.5, or any Scopus indexed journals without impact factor, even then 90% of them will straightway face rejection even before sending to peer review. But certainly, because of their highly digestive ethics these articles never get retracted or withdrawn even if an author publish twice or more the same content or plagiarise others work unethically.