I'm new to QTI and after having read the specs am having trouble implementing them in tests that i am creating by hand.

My assessmentItem consists in paragraph of text with multiple blank spaces that the user must fill.

Each blank space corresponds to exactly one correct result.

The outcome of the assessmentItem should be 1 if all results are correct and 0 otherwise.

I am trying to construct my assessmentItem with multiple responseDeclarations and a custom responseProcessor that checks that all are correct.

Is this the right way to do it?

Yes, that's exactly what you need to do

So for each textEntryInteraction, you make a response variable and check the value of that variable against the correctResponse value that you can set inside the responseDeclaration, e.g.:

Your response variables are declared like this:

<responseDeclaration identifier="RESPONSE1" cardinality="single" baseType="string">
<correctResponse>
<value>chocolate chip</value>
</correctResponse>
</responseDeclaration>

Then you have a chunk of text in the itemBody:

<p>The best kind of cookie is <textEntryInteraction expectedLength="20" responseIdentifier="RESPONSE1"/>, but ....</p>

And somewhere in the response processing you put:

<match>
<variable identifier="RESPONSE1"/>
<correct identifier="RESPONSE1"/>
</match>

and set the score and feedback suitably

<setOutcomeValue identifier="FEEDBACK">
<baseValue baseType="identifier">CORRECT</baseValue>
</setOutcomeValue>
<setOutcomeValue identifier="SCORE">
<baseValue baseType="float">1.0</baseValue>
</setOutcomeValue>

Remember SCORE should bea float variable.

HTH
Sue