If that is the case, why have this controversial patent clause at all?
I haven't read them all, but Facebook seems to have a number of patents which at first glance are so broad they could apply to react:
https://patents.google.com/patent/US20160092096A1/en "a method performed by one or more computing devices including defining a hierarchical structure for a user interface (UI) that includes defining one or more layers of the hierarchical structure, adding one or more objects at each layer, and specifying one or more relationships among particular objects"
https://patents.google.com/patent/US20160091953A1/en "Adjusting mobile device state based on user intentions and/or identity"
https://patents.google.com/patent/US20150277691A1/en "In one embodiment, as a user is scrolling through a first series of content items and reaches the nth content items from the first series of content items, display a visual indication that there are additional content items from the first series of content items existing after the nth content item."
https://patents.google.com/patent/US20150113066A1/en "A communications system including one or more alert gates and an alert controller. Each alert gate is configured to detect a different type of alert feed corresponding to a particular kind of alert. The alert controller is connected to the alert gates and operable to receive detected alerts from the alert gates and to deliver the detected alerts to a user of the communications system."
https://patents.google.com/patent/US20160086219A1/en "methods and systems that allows for improved user navigation within a group of content items."
I actually had no idea Facebook's patent portfolio was so broad. There are countless more: https://patents.google.com/?q=G06F3%2f048&assignee=Facebook+...
other than the unmitigated absurdity that is the patent system these days?
Even if I subscribed to the idea that this was without prior art on the web in 2009, equivalent functionality already existed in desktop applications. To claim that it is a novel invention because it now works on a web page is disheartening. It's exactly this kind of behavior--claiming new ownership based on each abstraction--that leads to restrictions in some countries on patents which can only exist in software.
I am not a lawyer either, but I think the idea of implicit grant is compatible with common sense.
> Wherever possible, make sure you have an explicit license to any necessary patents held by the licensor.
Then it goes on to distinguish between weak and strong grants.
I think you mean weak, not implicit. AFAICT, there isn't an implicit grant with patent law, because the default is that you're infringing without an explicit grant.
The theory behind implicit patent licenses is that the licensee (often the buyer) will have some idea about which rights they have to the thing they are licensing, absent any other agreement. For a software license like the 3-clause BSD, this means that if they are licensing the software from the holder of the patent, they can reasonably expect to not subsequently be sued for infringing on the patent while following the terms of the license.
Omitting any kind of patent mention from your product's license, and later claiming infringement for reasonably expected use, would look subversive and generally be unconvincing.
I'm not a lawyer. My understanding is that this is hypothetically possible but stands on shaky legal ground and has no legal precedent. Whether or not an implicit grant exists seems to depend on who you ask. This is apparently why GPLv3 includes an explicit grant.
It could also be that there aren't any patents that cover React. I don't personally know of any that do.