If you can write well (i.e., understandably and in a style consistent with the journal) it's not a problem at all. Just write your paper and submit it.

Based on my experience reviewing papers (many of which are from non-PhDs) I'd say that the most frequent problem I see is papers which were obviously written without having any thought given to the journal to which it would be submitted. Go read a few dozen papers from the journal to which you're submitting, and get a sense of what topics are covered in the journal (this doesn't necessarily match the list of fields mentioned on the journal's website), how much knowledge is assumed of the readers, and how formal/informal the writing style is; then write accordingly. Bonus points for downloading the LaTeX style sheet and formatting your paper properly -- we can talk about how style shouldn't affect things all day long, but the simple fact is that having your paper look "different" will distract reviewers from what you're saying.

You'll occasionally hear a PhD program being compared to an apprenticeship, and there's a lot of truth to that. Completing a PhD seems to be as much about learning how to orient yourself within scientific culture as it is about research or scholarship.

As others have said already, your field matters a lot. If you try to publish in a "softer" field, say history or economics, politics and outer markers count for a lot. If you are in mathematics, there should be zero problem if you are good. Alternatively, if you are in a field where success can be measured concretely, say in engineering, you should also have less difficulties. Good luck, whether you go for PhD or not!

One of those effects is the intellectual sponsorship of people in legit academia, particularly professors whom you have an IRL relationship with and can recommend you to the editor. If not in school, I think this only works if you directly build on research conducted by contributors of the journal in question.

I would think in CS this would be easier, because professionalism is young in CS. I think it would be an order of magnitude more difficult in the humanities, where accreditation is far more entrenched.

I don't know if other unis do this, but I would expect that it's part of the contract with journal publishers. Note: they don't publicise these terminals - just ask.

That said, good papers do share a common structure. Part of what you learn when you get a Ph.D. is what that structure is. My recommendation is to read through the previously accepted papers for the venue to which you're submitting. Note the level of exposition, formality, rigor, style of empirical evaluation, etc. When you submit, try to match that.

Non-academics fail most frequently on related work. Spend time researching related work and comparing it to your own before you submit! I've seen great papers dismissed out of hand, simply because the coverage of related work was deemed inadequate.

In computer science, many papers fail because they don't explain their core idea quickly or clearly. Failing to snag the reviewer's interest early on will lead to a grumpier reviewer, one more likely to magnify forgivable errors into unforgivable ones.

Lastly, if you get reject, take the feedback to heart, and keep trying. Academics have papers rejected all the time, and we learn early on to brush it off, take our criticism constructively and move on. Don't get discouraged!

That can be a very important consideration for a few reasons. First, if you're interested in publishing in the math field, forget peer reviewed journals entirely. From what I've heard, the arXiv (http://arxiv.org/) has essentially replaced most math journals. Similarly, while peer reviewed journals are still very important for physics, many papers will appear on the arXiv first (and in journals like Phys. Rev. Lett. later).

In theory, anyone can publish to the arXiv. In practice, they have a referral system designed to keep the crackpots out (with varying degrees of success). My recommendation would be to find someone who publishes there and is interested in the same topics as you, then contact them directly. No matter which route you pick, don't underestimate the value of personal contacts. As an alternative to the arXiv, Nature has a relatively new preprint server set up as well (http://precedings.nature.com/) though its popularity pales in comparison to the arXiv.

If math or physics is not your game (or even if it is), then I'd recommend PLoS One (http://www.plosone.org/). They have a policy of publishing any work with scientific merit. However, since they work on an Open Access model, they will charge you in order to review and publish your work. That said, if you can convince them that you don't have the means to pay their charge, they will sometimes wave the fee.

Finally, it never hurts to send off your manuscript to Science, Nature, or any one of the second or third tier journals. True, many of them won't give you even a cursory glance without degree and affiliation, but there's always a chance!

Submission is trivial: most have a webpage (with a deadline and submission requirements - e.g. double spaced, single column, no more than 12 pages), that you just put your document and details into.

That being said, most research areas are such that it's really quite difficult to position your work in such a way that it will get accepted: you need to know prior art well enough (and show this), but you also need to know who's work is actually any good (not easy) and who on the PC you need to reference, etc.

If you believe you can make a genuine contribution to a field, you need to pick a candidate publication, read it to understand its style convention, follow the "Instructions to Authors" religiously and pay the page charges (where appropriate). In your paper you need to link what you present to research that has been published before - think of science as case law - it builds up iteratively. By basing your work on previously peer-reviewed publications you are helping your reviewers see that you are working on solid ground.

Most editors will screen arrant nonsense ("I can prove Einstein was wrong"), but are likely to pass reasonable material to a review panel irrespective of any qualifications of the authors.

The real question is why you would want to. Generally being published in a peer-reviewed journal is only of real value to academics, because they are rated depending on their publication and citation rates. Otherwise, if you can anything interesting to say you can just publish it on a paper server (eg. arXiv as mentioned in another comment).

There's no shaggy-dog story of how we broke into the big leagues. We submitted the way everybody else does, and the anonymous peer review process never demanded our grad school union card. We had the benefit of amazing, experienced readers giving us unsparing feedback long before submitting, and had the good luck of having built an interesting system with some weird properties. Having something to say that your peers might find interesting is the hard part, not collecting qualifications.

There are basically two variables you're trying to optimize for: how closely your paper matches the journal subject, and how important your results are relative to the journal's average. The first point is easy to gauge on your own. The second is more difficult. You may not have an accurate judgment of how good your results are if you aren't intimately familiar with your field. It might be worth getting advice from someone who is. When you do know how good your paper is and how good different journals are, you can either overshoot (and hope for a better line on your resume) or undershoot (and decrease the chance that you'll have to wait through a second submission round).

[trying to shorten my post since it won't submit]

Finally, you should know that in CS, conferences are often preferred over journals. If there's an appropriate conference that you would be able to attend, you may want to consider submitting it to them. Disadvantages are paying for conference registration, advantages are a faster publication timeline and getting to attend and present at a conference.

Its like the newspaper industry which has outlived its business model.

Someone please youtube the whole thing and kill them already.

Ever heard of Google Scholar? As long as researchers put their papers up on their personal websites -- and almost everyone does, these days -- Google will find them. Sure, there isn't 100% coverage; but I'd say that Google Scholar provides much better coverage for published research than youtube provides for TV shows.

Maybe the situation is different in my fields (algorithms and cryptography) than in yours?

Interestingly the publishing industry has a firmer hold on the life sciences.

This is good too: http://math.ucr.edu/home/baez/journals.html

I liked the "Cite as" feature, but there are fundamental issues which I think needs to be changed:

1. Anyone should be able to comment/discuss articles. With perhaps karma based user voting to surface the best comments. Each comment should again be citable.

2. The text should be in HTML- shareable simply and acted upon by the community.

3. A reddit style list for each topic listing the recent,high impact papers. High impact is again voted on by the community, again with people who are 'known' to produce better research having more weight.

I don't think PhD is a pre-requisite or even makes much difference in peer review process. Of course, with a PhD you would have better experience and possibly better research. Many of the good journals remove names and adresses before sending for a review, so it is a blind peer review where holding a PhD degree or not does not make much difference.

In blind peer review this isn't a strict requirement, but it will help you put things in the right language and help you make sure that you're referencing the right things.