Re: [friam] Oracle trying to pollute the name "object capabilities"?

Given that linkers have been abusing the term "object" for decades and it
hasn't seemed to cause much harm to the more general meaning in computer
science, it seems to me that this very narrow-context definition of "object
capability" (which AFAICT applies to implementation details of the SPARC
ELF format that almost no one will ever see anyway) is unlikely to cause us
any harm. It doesn't even have anything to do with security. I don't think
this is comparable to POSIX or Netscape capabilities.

I think that calling this "a blatant attempt at false advertising -- or at
least parasitism and dishonesty" is likely to win only enemies, and make
people take us less seriously. It's obvious enough how they came up with
this terminology, given the context, and it seems likely the engineers in
question didn't even know about our definition of the term. They certainly
didn't intentionally steal the term.

IMO this isn't worth fighting.

-Kenton

--
Sandstorm.io is crowdfunding! http://igg.me/at/sandstorm


On Sat, Aug 2, 2014 at 4:04 PM, Mark S. Miller <[email protected]> wrote:

> On Sat, Aug 2, 2014 at 2:22 PM, Randall Smith <[email protected]>
> wrote:
>
>> Hi Mark —
>>
>> I understand your frustration here, sorry I don’t know to whom to address
>> your concern …
>>
>> ...but…
>>
>> Aren’t you equally upset by the use of the term “object” in this
>> document, denoting a chunk of memory that is not even encapsulated? I
>> gently suggest that the term “capability" is rather in the same situation
>> as “object.” Namely, it’s too routine, it’s too everyday … it’s expecting
>> too much for the word to acquire a universally-respected meaning more
>> subtly crafted than it has in it’s foggy everyday use. Even within a
>> sub-domain as broad as computational systems.
>>
>
> Hi Randy, I understand your point regarding "object" by itself and
> "capability" by itself. In both cases, I wish the term originally chosen
> for these concepts did not overlap with such everyday use of a natural
> language term whose meaning is broad and vague. However, I do not believe
> your point applies to the phrase "object capability". That phrase does not
> normally occur in natural language use. It was also wholly absent in comp
> sci until recently. In comp sci, it is our community alone that established
> it as a term of art, specifically because the term "capability" had gotten
> so corrupted.
>
> At this point, any use of the full phrase "object capability" in comp sci
> to mean something else is a blatant attempt at false advertising -- or at
> least parasitism and dishonesty. It is an attempt to inherit the positive
> connotations that the genuine concept has earned while providing something
> else that does not deserve those connotations.
>
>
>> With enough work, I could try to get you in touch with someone in the
>> right part of the Oracle org chart, but in order to have the most
>> constructive impact in the world, we need to select our battles wisely. In
>> other words, let’s not waste our time when we know we’ll just get the cold,
>> fishy-eyed stare.
>>
>
> Please. Indeed, I have very little time for such battles, so I pick very
> few. But I'm willing to bet my scarce battle time on the notion that this
> one is worth picking.
>
>
>
>>
>> FWIW, in my opinion, you’re under no threat of concept dilution. Anyone
>> learning about capability systems will immediately appreciate the
>> expressive power and elegance they offer, and, as with “object,” come to
>> live their now-enlightened life without confusion when encountering some
>> document that wants to use the word in a different way. (An Oracle document
>> no less! Come on! I mean… an Oracle document!)
>>
>>
> One of the most pernicious corruptions that originally drove us to seek a
> new name was the use of "capability" to name so-called "Posix/Linux
> capabilities". In this case, the dishonest intentions of those that chose
> the name are clear -- they cannot claim ignorance. However, by the time we
> might have considered fighting that battle, it was already too late -- they
> were already entrenched. The fact that Oracle's misuse of "object
> capabilities" is not yet entrenched is exactly why this battle might still
> be winnable.
>
> OTOH, supporting your point, we also fought so-called "Netscape
> capabilities", which was actually part of their very non-capability stack
> introspection architecture. We also couldn't get them to change. However,
> they disappeared on their own so, in retrospect, the effort we spent
> fighting them was wasted.
>
> By the time some terminological abuse has grown to the point that we can
> know it was worth fighting to kill it, it is probably already too late to
> do so.
>
>
>
>
>>     -- Randy
>>
>> On August 2, 2014 at 10:57:42 AM, Mark S. Miller ([email protected])
>> wrote:
>>
>>  [+...]
>>
>> DAMN!!!
>>
>> Who should we complain to in order to get this fixed?
>>
>>
>>
>> On Sat, Aug 2, 2014 at 9:38 AM, Mike Stay <[email protected]> wrote:
>>
>>> http://docs.oracle.com/cd/E23824_01/html/819-0690/chapter7-28.html
>>>
>>> --
>>> Mike Stay - [email protected]
>>> http://www.cs.auckland.ac.nz/~mike
>>> http://reperiendi.wordpress.com
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "friam" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to friam+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
>>> To post to this group, send email to friam-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
>>> Visit this group at http://groups.google.com/group/friam.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
>>
>>
>> --
>>     Cheers,
>>     --MarkM
>>
>>
>
>
> --
>     Cheers,
>     --MarkM
>
> _______________________________________________
> cap-talk mailing list
> [email protected]
> http://www.eros-os.org/mailman/listinfo/cap-talk
>
>

_______________________________________________
cap-talk mailing list
[email protected]
http://www.eros-os.org/mailman/listinfo/cap-talk