Not being a native speaker of English does of course affect your ability to publish in English speaking journals, but it cuts both ways.
Many journals and conferences these days are dominated by non-English speakers, while the majority have a degree of balance. The transactions style journals in engineering and computing publish shorter papers with a more mathematical flavour and a conventional format, and often have editors and reviewers with an Asian background. In round numbers less than a quarter of the world's population speak/write English with native level efficiency, the major languages/grouping being English, Spanish, Chinese and Indian - and their dialects and relatives.
As a native speaker of English and a linguist, I have sometimes been told by reviewers that I need to improve my English, include more display equations and more references to their journals, and get my paper checked by a native speaker! I can quite easily see their linguistic background from the idiosyncrasies and errors in their review. Although this may sound laughable, there is a message here for both of us, the native speaker and the non-native speaker of English.
As a native speaker and an academic used to using English in a sophisticated interdisciplinary way, I need to learn to write in a simpler way to be comprehensible to readers who do not share this background - whether native English speakers or not. For the non-native speakers, they need to do exactly the same thing. In fact, computers also need simpler language, so for natural language processing, optical character recognition, speech recognition and machine translation systems, the same simplifications will also help.
This simplification includes
- reducing sentence length and complexity
- reducing vocabulary size and complexity
- reducing the use of jargon and idioms
- reducing the use of informal models
Let's explain this a bit more, and I'll draw on some Computational Linguistics background to help here... Parsers can't handle long sentences and shortly before the turn of the century people started to scrap sentences longer than 40 words before publishing statistics about their systems - because they just couldn't handle the complexity and ambiguity. English is very rich, and individual words and phrases can have many meanings and alternative apparent synonyms can have very different nuances. This creates a lattice of possibilities. In English every second word tends to be a simple one, one of the 150 most frequent words, and to have a grammatical function, but often overloaded with different possible meanings (at least two on average). These more grammatical words are called functional closed class words (there is a fixed number), and the others that have more meaning associated with them are called contentive open class words (there are always new words and usages being invented). Many words have a dozen or more meanings (but again assume at least two).
So a 40 word sentence can have a million (2^20) legal grammatical parses and a million (2^20) semantic combinations expressed as a network or lattice of possibilities to choose from. We choose the intended grammar and meaning based on the salience of the concepts in our physical or linguistic context - in the real world, that includes things we know and see and hear in our society/environment, but in a paper it includes the things an expert in the field should know and the ideas expressed in the sentences or paragraphs before. When speaking, we see when someone isn't following, we hear when something is ambiguous, and we adjust accordingly. We might stop a sentence half way through and go back to basics. We might pause and repair with a parenthetic 'that is'. We might simply say 'pun not intended' because we've recognized that there is another meaning that jumps out.
In text we need to anticipate how different readers will understand things, what the ambiguities are, where the complexity lies.
In relation to length, if a sentence is more than 40 words, or 4 lines, chop it up. I tend to make use of participles (-ed and -ing) words and relational connectives (which/that) to join phrases and clauses together. I tend to put in lots of parenthetic comments, often shown with commas rather than parentheses, but these add complexity. So look particularly at the wh-, th-, -ed and -ing words (closed class functional/grammatical words or suffixes that can be used in many different ways). Break these up into separate sentences and that way the non-native reader or writer won't get confused by the subtleties of how they are used. Words like 'to' and 'through' and 'that' can be a problem, and reflect different usages between different brands of English. Sometimes the problem is that a th- word is omitted, particularly 'that' or 'the'. Adding the word in can resolve ungrammatical or ambiguous sentences.
If a paragraph is more than about 10 lines, chop it up. People scan papers quickly with their eyes being drawn to the headings and the first sentence and last few words of each paragraph. Make sure paragraphing is clear and the formatting compact (modern journals don't require double spacing which is designed to slow the proof reader down, and should not be supplied to people are meant to be understanding the content, and inline use of equations or superscripts should not cause line spacing to open and close like an accordion - select fixed mode and make sure everything fits). The first sentence of a paragraph should introduce the idea explained or developed. The last should provide a bridge from that idea to the next. A punchy paragraph may have a single line, particularly the conclusion at the end of a section. A long paragraph may consist of three or four sentences.
Note. Things you have in brackets that make a sentence or paragraph too long can be moved to a note. If you are using (hyperlinked) footnotes or endnotes, yes it can be that sort of a note, but endnotes are archaic and don't really help due to the need to flip back and forth. The comment can simply be moved to a following paragraph.
In relation to content word usage, make sure technical terms and acronyms are defined at first use in the body of the paper, and acronyms are refreshed at the beginning of each section. Definitions of mathematical variables may also need refreshing, particularly around equations or in table or figure captions. That is people forget these things, or skipped over them without taking them on board. They should be able to look at a table or a figure or a section on their own, and make sense of them without (re)reading the preceeding text. Avoid informal speech forms and idioms (phrases that have special meaning to people of particular nationalities rather than following from the standard meanings of the words used). Language teachers like to teach idiomatic phrases to their students to make them think they sound more natural, but it has the opposite effect because they invariably use them inappropriately. Use the appropriate technical terms, but explain them in simple language if your whole audience can't be expected to be familiar with them. Be sure you are using a word of phrase correctly before you use it. Don't just copy things other people have said, particularly authors who are clearly not native English speakers.
In relation to functional word usage, be careful of words like 'this/that', 'the/a'. English normally requires an article or a conjunction or relative pronoun, and while the native speaker knows when to omit them, it is normally grammatical to include one. Also be careful of words with negative connotations, the word 'not' and hedges like 'hardly' or 'quite' - it is easy to miss such a word when scanning quickly, and get the wrong impression, and non-native speakers tend to use the hedges incorrectly. This again is something computers are bad at, and information retrieval, recommender and ranking systems can't currently deal with such words correctly.
It is also important to talk about equations, particular from papers by non-native authors, or in papers sent to non-native editors/reviewers. Don't just copy equations from paper to paper. Always go back to the original source, the originator of the idea, to understand equations, models and assumptions. Always explain the equations and the insights that lie behind them. Many authors/journals use models and equations in a way that obscures their meaning and makes it hard to understand. The non-native speaker is particularly tempted to avoid trying to explain something in English by expressing it in Mathematics. Many students (and all too many academics who should know better) copy equations into their papers that they don't understand (which are thus often wrong, incorrectly applied, or simply inappropriate to the paper). Mathematics should be used to clarify not to obscure. Explanations should always be provided. Don't describe what going on in operation-by-operation words, but explain the insights and the effect of the equations.