performance of something vs. the performance of something

Question

I asked my American friend to take a look at these three sentences and pay particular attention to how the word performance was used. To him, they all sounded fine. But what bugs me personally about the grammar in all three sentences is that in the first two examples we don't need a definite article in front of the word performance while example #3 requires that there be one in order to make the sentence sound grammatically smooth. Do you have any explanation as to why that is the case?

Example #1:

Performance of thread-related features—and particularly synchronization constructs—is key to writing multithreaded programs.

Example #2:

Performance of applications that tend to use input/output capabilities extensively during their execution can degrade significantly especially when run in networking environments such as local area networks.

Example #3:

This strategy will really help you make the performance of your applications much better.

Answer 1

Performance (of something) is an abstract noncount noun. It is abstract because you cannot physically touch it, and it is non-count because you cannot have "12 performances".

Well, an artist can have 12 performances, but that would be another sense of this lexeme. If an artist holds 3 or 4 performances in a row, that might affect his performance.

Other non-count abstract nouns are kindness, employment, happiness, honesty, literature, history, sleep, etc. Some of them can switch between count and noncount sense (she showed me many kindnesses; the review compares many histories of post-Soviet Russia).

When an abstract noun is used in its non-count sense, it can be used generically: used to refer to some notion in general, in the widest possible sense. If such a noun is not modified (has no adjective or a prepositional phrase attached to it) we usually omit the:

1. She is studying history. ("History in general": the noun is not modified)

When such a noun is pre-modified, we also usually omit the:

2. She is studying Russian history. ("Russian history in general": the noun is premodified by the word "Russian")

On the other hand, when such a noun is post-modified, it is usually preceded by the definite article:

3. She is studying the history of Russia. (post-modified by the prepositional phrase "of Russia").

According to Quirk et al., the last sentence slightly differs in meaning: while in (2) we are sure she is studying Russian history as a whole, in (3) it is possible that she is studying just some aspects of Russian history, or a particular university course.

Since noncount abstract nouns postmodified by "of-phrases" usually take the definite article, it is understandable that one feels the urge to append the before your sentences 1 and 2:

The performance of thread-related features—and particularly synchronization constructs—is key to writing multithreaded programs.

The performance of applications that tend to use input/output capabilities extensively during their execution can degrade significantly especially when run in networking environments such as local area networks.

But since according to Quirk et al. that might open the door for re-interpretation - this is just my guess as a non-native speaker - hence possibly in some technical contexts people avoid using the just to make sure that the phrase is interpreted generically - in the broadest possible sense.

Let a native speaker confirm or refute my conclusion. I wrote the answer mostly to refresh my memory, because this issue plagues me as well..

---

Reference:
Quirk et al., "A Comprehensive Grammar of the English Language", 5.58, The articles with abstract noncount nouns.

Answer 2

To me, I think you need "The" definite article on the front of the first two, as you are speaking about 'the' performance specific to or of thread-related features and applications.

I don't see any risk of ambiguity, even though you are speaking generically and not specifying any one or a specific number of thread-related features or applications. As I understand it, you are speaking about the performance of a single but unidentified member of the class 'thread-related Feature' or 'Application'.