Brendan Eich is a user on mastodon.social. You can follow them or interact with them if you have an account anywhere in the fediverse. If you don't, you can sign up here.
40cd7d5f03866bed

Brendan Eich @BrendanEich

Brendan Eich boosted

@BrendanEich Spent an hour this morning with an [insurance industry] client setting up a load of write-only reports. It ain't just banks.

Brendan Eich boosted

Oooh. Interesting.

Per twitter.com/doctorow/status/86 at birbsite, certain tethering blocks are defeatable by changing MAX_MTU

I wonder what this will do to other kinds of packet shaping and QoS schemes.

@munin Ok, it was a bank. Could you tell? And an allegedly high-tech one, at that.

Hilarious escapade doing online application at great pain (form validation bugs, inconsistent state in app UX vs. data model), only to find after traveling that ink-on-paper signature needed. Printed whole thing, signed. Hotel-provided Epson could not scan, it used 10-year old "USB" tech. Front desk scanned whole double-sided, ink-on-paper signed printout, sent me two PDFs: first odd pages, then reverse order even pages.

Institution requiring this is happy now, even with the split/reverse, lol.

Brendan Eich boosted

This is why I am against biometrics as a sole identifier:

theguardian.com/business/2017/

There's no biometric sign that is uncloneable. It's OK as a username-equivalent, but it is unforgivably bad as an authenticator.

Brendan Eich boosted
Brendan Eich boosted

@RickByers @Rushyo How would you fix it? The degree and kind of compatibility required to compete is not determined by the browser vendors alone, but the set of sites users prefer to visit, and their owners paying for maintenance and upgrades. We are seeing more "break the web" now, e.g. WICG with Chrome (yourself on point ;-) leading, or ad/tracker blocking by default as in Brave. But it is hard for other than market leader and small share browsers to do this.

@metapunk @Rushyo @fred Not knowing all the constraints, like everything else we don't know, favors incrementalism. So Quantum not Servo in full. Even in best case, it requires time: multiple years.

@fred @Rushyo @metapunk MMM definitely applies. Also design is hard, coding even in Rust hard. Worse, not all the constraints are known. Best if Servo can aim for future Web where, after WICG "interventions" and rise of ad/tracker blocking, it can avoid old, now-obsolescent-yet-still-obligate-in-2017 constraints.

@fred @Rushyo @metapunk If doing an engine with its own legacy constraints, the cost is high. If doing a browser that innovates above the engine, then easier. The problem for Firefox is changing the engine while the plane is in the air. Quantum is the right direction but it'll take many release cycles.

@RickByers @Rushyo How would you fix it? The degree and kind of compatibility required to compete is not determined by the browser vendors alone, but the set of sites users prefer to visit, and their owners paying for maintenance and upgrades. We are seeing more "break the web" now, e.g. WICG with Chrome (yourself on point ;-) leading, or ad/tracker blocking by default as in Brave. But it is hard for other than market leader and small share browsers to do this.

@cnc I'm dodging nothing and I don't owe you my time. Your "heads I win tails you lose" approach is tiresome. Muting.

@cnc I don't think you know the first thing about me. I've written over the last few years on HN if you care to read. One entry point: news.ycombinator.com/item?id=1. If you can't engage other than by rigid assertion of hate and demands for me to atone, I'm afraid there won't be any point in talking here.

@cnc I think you misspelled "have" - I had a reply typed about how Brave in fact employs valued people of diverse backgrounds, but your delete caused Mastodon to eat it.

That was a much better question than your revised Senator-when-did-you-stop-beating-your-wife one. Please! This is Mastodon, not birdsite.

@sajith If only I had proper long-form time to spare ;-). I'll do it as I can.

@Rushyo @metapunk Firefox with paid crew of 10 (11 in 2004) had significant help from volunteers whom AOL laid off in July 2003, many on severance packages. We contracted & hired some of these folks as search revenue from Google came in, starting in August 2004.

Trolltech supported KHTML if not Konqueror folks (I knew some of the former) which gave a boost to Safari, leading to webkit.org.

I still can't think of ~all-volunteer/large-scale/long-lived examples. GCC? LLVM counterexample.

@metapunk @Rushyo on your last point, siliconangle.com/blog/2017/05/.

Netscape/AOL supported mozilla.org for 5 years while losing the remaining market share held by its browser. That cost big $ and was crucial to our ability to do Firefox (started in 2002, downsized crew of 10 paid in 2003, 1.0 in Nov. 2004).

In rare cases, I've seen open source "go slow" or even hibernate on nearly all-volunteer life support, then come back.

@jansegers Thanks for the link, I had not heard of IEML. My friend jg from SGI and Netcape founded Metaweb with Danny Hillis and sold it to Google. Metaweb's Freebase. Lots of hand curation, some ingesting of structured data from wikipedia. I'll have a look at IEML when I have time -- thanks again.

@Rushyo The indie / consulting model can be good from what I see, of course has to avoid getting captured by one big client. I've never seen it raise funds to the level of Mozilla's engineering + minimal docs/marketing/PKI-policy-etc. budget.

@Rushyo Opera's Presto never had enough desktop share to get web dev testing, so they switched to chromium/Blink in 2013. If Firefox share falls toward Opera desktop levels, Gecko is at risk of same effect.

There are AFAIK very few FOSS projects of the MLOC-measured scale of Gecko+Firefox, never mind Rust and Servo, etc., that are funded without large / single sources/sponsors. GCC comes to mind -- it had multiple companies supporting it over its life to date. Can you cite an example? Thanks.