LanguageTool has entered feature freeze because it will be released next week.
However, I would like to add one last rule before the release (and of course, continue fixing the gender antipatterns).
Take a look at these two sentences:
O consultor redigiu um relatório longo que não vontade para ler.
O consultor redigiu um relatório longo que vontade para ler.
It is missing a verb before “vontade”.
So, I was planning to create a rule saying that there seems to be missing a ligature verb.
Do you have any idea of the name for rule id, rule name, suggestion message?
Also, what is your opinion regards the tokens to check it?:
longo? + que + [não] + vontade + para + something.
should I directly use “que” and “para” in the rule, or use POSes?
Also, if there is no negation (“não”) will the rule cause false positives?
Jaume made some changes in the disambiguator and I will code/fix rules after tonight’s nightly, which means doing it at 5am tomorrow.
Hi,
Before going into details of names, IDs, let’s consider some cases:
Ele disse que vontade não falta.
This sentence is perfect. How will the rule differentiate it from the one it should apply?
Also:
Ele disse que não vontade de escrever.
Suggesting a verb here is tricky: it maybe tem, teve, tinha, há, houve, and if it is referring to some people, as in Ele disse que não temos vontade de escever, the range of options increases.
I can’t see a way to go away with these problems, but if you have any idea, the rule may have a life!
What I was trying to say is that the rule would just mention the necessity of a missing verb, but it wouldn’t say the verb name, just mention that a verb could be missing.
Hi, sorry for the delay in replying. The sentence:
Ele disse que vontade não falta.
is similar to your examples:
O consultor redigiu um relatório longo que vontade para ler.
I’m afraid that it will be hard to solve how to apply the rule in the second case, but not in the first.
But it may be worth a try: maybe sentences like “Ele disse que vontade não falta” are rare and don’t appear in the corpus, so the rule would be fine.