Umm, are you just asking for a simple replace? As in something like s/Lorem ipsum blah blah blah/Salsa/? Or are you asking something more complex?

A regexp doesn't reduce things, though. It just matches (or doesn't match) them. It has no output capabilities.

Maybe I'm misunderstanding you, but this seems like it'd be fairly simple if tedious task, for which is possible to prove both the existence of a solution, and easily provide an algorithm to generate it(if you dont want to just muck around with regexp syntax which also seems pretty straight forward). You want to build a regexp to recognize the given string. This is equivalent to wondering if you can build a DFA that can recognize the string, which clearly you can, as for example below

state0(on l go to state one, else computation fails)
State1(on o go to state 2, else computation fails)
etc..
StateLast (we recognized the string, output salsa)

Really since you are only trying to recognize the one string, theres not much to it.

Then you want to convert the DFA to a regular expression. For every DFA there is a regular expression and algorithms exist for converting one to the other..So really not too exciting, although tedious to apply to the long state machine that would result.

Perhaps I'm not understand what you mean by reduce, but imo any reasonable definition you can come up for that would just mean outputting salsa whenever you recognize that string(or outputting parts of salsa when you recognize parts of the string, but thats to be equivalent) .

you are asking for a regexp that would match only one specific text?
(in formal words: a language of only one word)

that is not a challenge.
the regexp is ^text$

a challenge would be to identify any text that an average human would confuse for the lorem ipsum :)

This is the closest thing to "reducing" that text to "salsa" in Perl. (There is no uppercase "S" in the text, that's why not "Salsa".




But this problem is not challenging nor interesting at all.