I am teaching English in a Japanese primary school and they have a Nepalese child who doesn't understand written Japanese especially so they got me to translate some maths problems for them. I had a lot of those sentence type maths growing in Australia and I wondered what the point of writing sentences that need to parsed by students to extract the mathematics is.
I read somewhere that they think it's real although there would be situations where someone would write a sentence on a situation just to be mathematically analyzed. I can't think of an engineering (my first degree) mathematics situation like that. For instance, in calculus there is no way to have a true sentence problem---it would take many paragraphs to express the many complicated expressions and that would be very difficult for a university student to understand; they don't have that enough time in exams. It's not the same as having an engineering problem to solve---not that we did many.
There are many ways to include context into a maths question (without using sentences with lots of extraneous detail)---diagrams, tables, short phrases (e.g., 240V +-15, 60Hz according a standard)---so that it isn't that hard to linguistically interpret. An applied math problem is not the same as a pure math problem with the same nominal mathematics. For example, $2+2=4$ but the applied maths example is you need $2$ lots of $2$ m of steel and the only available lengths of steel don't include $4$ metre lengths rather only $5$ metre lengths then it's effectively $2+2=5$ (+extra $1$ metre) in metres.
You also have to differentiate between reading and writing. It's a bit easier to write than read---you understand your own writing more and writing summary statements is not what the question is about.
Sure there are some documents in some fields where there are numbers but it's unclear whether pushing primary school children with the simplistic sentence type problems that are normal helps them do such problems. Please explain the value of sentence type problems as against other contextual forms for preparing young children for possible real problems.
Can anyone spot what's wrong with the Paint problem below?
$\frac56$ dL ($1$ dL $=0.1$ L) of paint covers $\frac34$ m$^2$. How far does $1$ dL of paint cover?
It should have shown a paint can and the painting done from it.
What's wrong with the Steel problem below?
$\frac27$ m of steel bar weighs $\frac45$ kg. How much does $1$ m of steel bar weigh?
It should have had a metre length of steel and its weight with an arrow.
What is wrong with the answer key (teacher's copy is shown) is that it doesn't include any details of the context except units.
What is wrong with the context is that primary school students wouldn't have any idea about paint economics or steel usage. These are totally foreign to them and so students know to skim out the numbers. But this is what happens when we insist on having sentence problems---they have to be artificial and superficial otherwise the effort of teaching fractions is very diluted with teaching context and sentence structure.