That would not be an example of the relationship "correct", since you're posing the question without a specific theoretical context, e.g. "Blochian phonemics" vs. "Generative phonology". The classic and standard (though not historically earliest) taxonomic understanding of the object of study is that "phoneme" is a grouping relation on surface phones, where if phoneme /X/ is defined as the set [w,x,y] then the content of /X/ is all and only the instances of [w], [x] or [y]. It is a requirement that the distribution of [w,x,y] not overlap, w.r.t. context that they appear in. Since the cited usage of "allophone" does not respect that condition, that would not be a correct application of the terminology.
However, I don't think the above-mentioned classical characterization has been actually believed in for a number of decades, and the page does correctly reflect the more modern understanding of phoneme and allophone where a "phoneme" is an item that can appear in underlying forms, and an allophone is an item that can only appear in surface forms.
When there are competing definitions of terms, it would be misleading to say that a given use of the terminology is "correct", if it is only correct assuming one definition. If a usage is at odds with all definitions then we can say that it is incorrect, and if a usage turns out to be consistent with all definitions then we can say that it is correct. In the present mixed-bag situation, the most you can say is that it is possible -- depending on what definition you're assuming.
The reason why people point back to the definitions of the 40's is that they represent the pinnacle of axiomatic, formal thinking on the topic.