To get a preliminary injunction, Apple needed to convince the judge that it is likely to win at trial and also that it will be irreparably harmed if Amazon is not blocked from using the mark during the time leading up to trial. The judge said no to the preliminary injunction, not on grounds of lack of irreparable harm (that issue was never reached) but on the ground that Apple was not likely to win on the merits of its claims at trial.
Marks are classified as generic, descriptive, suggestive, arbitrary, or fanciful. Any mark that is suggestive, arbitrary, or fanciful is protectable. A generic mark is never protectable and a descriptive mark normally is not either - but the latter can become protectable if it acquires distinctiveness in the public mind through what is called "secondary meaning" (which generally means that a company has advertised a product or service to such a degree that the public comes to associate the mark with the producer of the goods or services as opposed to seeing the mark as descriptive, e.g., "Holiday Inn").
Apple claimed that the "app store" mark is suggestive because its meaning is "not inherently obvious" and because Apple was required to devote a lot of resources to educate consumers about the service.
Amazon claimed the mark is generic in merely describing a store where software apps are sold.
On this key issue, the court ruled as follows in assessing whether Apple was likely to succeed with its claim: "The court assumes without deciding that the 'App Store' mark is protectable as a descriptive mark that has arguably acquired secondary meaning. The court does not agree with Amazon that the mark is purely generic . . . but also does not find that Apple has shown that the mark is suggestive, as there appears to be no need for a leap of imagination to understand what the term means."
So far so good for Apple (with its major advertising, it may well be able to prove secondary meaning at trial).
But Apple got sunk by the second factor it will need to show if it is to win at trial and that is the "likelihood of confusion" factor. On this point, the court weighed eight factors as required by governing precedent but, in effect, really boiled it down to the fact that consumers buying apps for Android devices will not really believe that they are buying them from Apple or from an Apple-authorized source when they see the term "app store" at Amazon's site.
The court also rejected Apple's dilution claim as a basis for granting preliminary relief, basically finding that "App Store" did not constitute a "famous mark" on the evidence shown by Apple to date (it is here that the generic references to "app store" by Mr. Jobs and others come back to bite Apple).
All this suggests that Apple will have an uphill fight at trial on this one (as it should, in my view).
The 18-page decision is well-reasoned and may be found here: http://www.scribd.com/doc/59494852/Apple-v-Amazon-Order-Deny.... Also, my prior comments on the case: http://news.ycombinator.com/item?id=2351910.