The title field, by convention, indicates the question being asked. The body field may also contain questions, or elucidation on the question. However, many questions have titles that are not questions, and therefore rely on the body to be clear what question is being asked.
The implied follow-on question is 'should titles that do not contain questions be edited?'
Ideally, yes.
When titles are questions, they're more likely to actually describe the post. Looking at the related pane of this very question will demonstrate that.
However, there are cases in which it is difficult to summarise the post into a single question, even if the question is narrow enough for a single answer. As such, it might be acceptable to post a question with a non-question title and wait for somebody else to edit the title into a question.
As always, if you think you can improve a question, edit it! That's what the feature is there for.
Wizzwizz4's answer already brings the main point home: Ideally, yes
Having tried myself to improve less than optimal titles, I had to learn that it's not as easy as assumed. A titel should be a question, but it needs as well to give a rough hint on the backgound. Forming it as a question may come second.
The most importent part of a title is being a title
A title's purpose is to help the reader. It is supposed to give an idea what the following text is about, as well as giving an introduction to the text following. When already opened, it's easy to continue reading from a less then perfect title into the question text.
This can't be done when questions are presented as title only. Like listed on the front page or when searching for a tag or term. Here the Title should, most of all, help in selecting a question - without the need to open and read the full text. Content will rule over form - turning the title into a 'right' question may be damaging. After all, this is RC.SE, not Jeopardy.
So turning 'Coding a Jump' into a grammatically correct question like 'How to code a jump?' may be nice, but next to worthless in real life. Changing it (based on the further content) into 'How are jumps coded in /370 assembly?' in contrast tells quite a lot and enables reader to select (or ignore) it at title level instead of looking into the text.
Related (though not asked), there is as well the danger to narrow down (or widening up) a question beyond what the OP intended to ask, just because one thinks this needs to be seen in a wider context (or more specific). It is not the task of an editor to change meaning no matter how tempting it is. Above any improvement in grammer or wording it must be the intent to keep the integrity of the original question. If in doubt, ask in a comment and wait for the OP to respond is the only way to go. Never edit without. Better leave a question unedited than making a distorting change.
In most cases, when it's understood that a question needs to be widened (or narrowed), the right way is to do so in an Answer. Answer what has been ask as good as possible and then add a paragraph as suffix (or prefix if must) stating why it needs to be widened and then add a more inclusive answer.
Albeit, if it is felt that the wider/narrower question is needed to be asked: Just ask it - possibly with a paragraph added why it isn't a duplicate, but a different case.
In general I found the process of thinking about why it needs to be different always lead to a more focused presentation - or, more often than not, the conclusion it's not as important or relevant as thought in the first moment.
There is a long-standing question on this on the main meta - Should "questions" be questions?
The consensus is "usually yes".
It is possible to construct a coherent question title that is not, in itself, an interrogative. However, for clarity and for ease of finding in searches, a question title is usually best when posed as a question.
