That curve is simply representative of a certain set of conditions;
Use something besides Ilfotec HC developer, change the dilution, modify the agitation scheme, adjust temperature, or adjust any combination of the factors and you get a different curve, each one unique.
Ilford's ID-11 curve actually shows to shoulder sooner.
The published curves I've seen all seem to fit in a given box, they show the toe and what I might call the normal exposure print range. Kodak is a bit better than Ilford about showing alternative condition curves but they still don't show what happens farther to the right. We simply have no info from Ilford about what happens with more exposure.
If you can show me an hp5+ curve with 14 stops in the linear region then I will concede you this point. Of course there are different curves, but none of them show 14 stops. If you still think it is possible then I invite you to tell me how, rather than simply claim that it isn't impossible.
Regarding the actual exposure range of the "straight" line, I'll have to agree I've never seen 14 stops in any film I've ever tried with any normal CI process. Acros would probably come closest. That's without flare. Under real shooting conditions the straight line would tend to be shorter due to compression at the low end.
Having said all that, I think PeterB may be a little too focused on the straight line. Depending on the film and process, there is usually a range of still very satisfactory local contrast beyond the straight line proper. For example on the upper end in my experience there are usually at least 2 stops at the begining of the shoulder which, for all intents and purposes, have "full" separations.
We're ultimately concerned with print quality (I hope), so we need to keep the film curve in perspective. If you didn't have curves in front of you, you probably wouldn't notice the difference between the straight line and early shoulder while you were printing your highlights. In addition, there is never any guarantee the full range of your shot ends up where you expect it to on the curve anyway. There are too many variables in the field (metering, exposure decision, flare, etc.).
I wasn't suggesting that HP5 would have 14, in fact the example I used actually showed a shoulder that started earlier than what you planned on. I was simply suggesting that the number isn't a known.