roca

71 ms·

781.
by roca 6y ago
There isn't any discussion in the document of why this effort will succeed when previous efforts along very similar lines failed.
782.
by roca 6y ago
That quotes a glibc comment "we also support negative `signed char' values for broken old programs" which means the OP code is in fact going to behave OK with glibc, providing that "support for broken old programs"
783.
by roca 6y ago
What exactly do you mean by "schedule-based safety model"?
784.
by roca 6y ago
The code for that blog post is fragile in an interesting way. Its safety and security depend on a subtle invariant: that the 'word_regex' does not match a string containing any non-ASCII character. If requirements changed so the r
785.
by roca 6y ago
Unfortunately that quote is an all-purpose way to dismiss the volume of complaints about any language.
786.
by roca 6y ago
OK, but that only works for prologue/epilogue --- it's better if you can use the instructions elsewhere too.
787.
by roca 6y ago
My gripe with LOOP and other crappy instructions is that they use up valuable space in the instruction encoding.
788.
by roca 6y ago
Limiting the registers to be specific pairs is awful if you're implementing a register allocator.
789.
by roca 6y ago
Tragically LOOP is alive and well in x86-64.
790.
by roca 6y ago
Thanks, I stand corrected; he did mention the Internet. At that time, he was still expecting a walled-garden AOL-like model to win the day, and Microsoft was frantically trying to build MSN to take that market away from AOL, so he downpla
791.
by roca 6y ago
> Yuan migrated to the USA from China in the 80s. He had heard Bill Gates speak about the internet The same Bill Gates who neglected to mention the Internet in "The Road Ahead", published 1995?
792.
by roca 6y ago
I always thought unidirectional type inference (i.e. C++ 'auto') was good enough but I have been surprised to find omnidirectional type inference very useful in Rust... for example, constructing an empty collection to pass as the
793.
by roca 6y ago
OK, thanks, you have a fair point. I still believe that defining "herd immunity" as just barely reaching the HIT will confuse readers of the declaration, who will have mostly encountered the term in contexts where it means the dis
794.
by roca 6y ago
That paper certainly does not admit their definition. The abstract is very clear: > A common implication of the term is that the risk of infection among susceptible individuals in a population is reduced by the presence and proximity of
795.
by roca 6y ago
My parents are 80, currently in good health, but definitely at high risk if they caught COVID. They're well-off and like to go out and spend money. Any kind of "focused protection" strategy means they have to stop going out a
796.
by roca 6y ago
It's worrying that these supposed professionals don't even know what "herd immunity" means. From the declaration: > herd immunity – i.e. the point at which the rate of new infections is stable This is not the correct
797.
by roca 6y ago
Another interesting question is, what is the economic impact of those high risk people not being able to leave their homes? What happens to the schools and workplaces where they normally serve? > A comprehensive and detailed list of meas
798.
by roca 6y ago
That is interesting. Unfortunately from my point of view it's hard to know what to make of a report like that at this late date.
799.
by roca 6y ago
I agree that mistakes were made. It is at best unclear that better rollout would have made a tangible difference to the outcomes.
800.
by roca 6y ago
Yes, I replied to your comment in which you linked to such an article. That is the "only Mozilla group that I know of" I referred to. It is worth noting (again) that those people were employees of the Foundation, not the Corporati
801.
by roca 6y ago
I'm not here to justify Mitchell's pay. But you can't write an article about the decline of Firefox market share during the rise of Chrome and blame it all on the CEO. (Mitchell's only been CEO for a year or so, but it i
802.
by roca 6y ago
Yes. FirefoxOS had some obvious revenue potential. The pattern here of course is that every attempt to raise extra revenue caused storms of protest amongst Mozilla supporters.
803.
by roca 6y ago
Regarding speed: it's not that Firefox got slower, it's that Chrome set a new performance bar that Firefox had to meet --- and largely has met! Though a cost of meeting that performance bar was dropping XUL extensions, which is on
804.
by roca 6y ago
Fair point. I've updated my HN profile. I'm Robert O'Callahan, formerly a Mozilla Distinguished Engineer. I wasn't on the board, but my claims are not about the board.
805.
by roca 6y ago
Everyone on that list worked for the Mozilla Foundation, and would not have reported to Brendan as CEO of the Mozilla Corporation (which made that tweet misleading at best). I referred to that group already in the comments here. AFAIK no Mo
806.
by roca 6y ago
KaiOS shipped 100M units by May 2019, no idea what the number is now but it's much higher. Not quite "fundamentally irrelevant".
807.
by roca 6y ago
This article has many problems. One of the main problems is that the immediate reason for Firefox losing market share is totally outside the control of the CEO or anyone else at Mozilla: Google pouring billions into Chrome, both actual bill
808.
by roca 6y ago
No, it's hard to get. Also tricky is what counts as "the engine" varies from project to project. E.g. Webkit doesn't cover as much functionality as Gecko or Chromium, e.g. the HTTP stack is separate.
809.
by roca 6y ago
If that "fork" was to survive as an independent browser engine with competitive performance and features you'd need at least 10x that number of developers.
810.
by roca 6y ago
The average Mozilla Gecko engine developer can quit and get a job at a big tech company making a lot more than $200K a year, definitely if they live in the Bay Area, and probably even if they don't.

More ›