After some tweaks, including enabling jumbo frames, I’ve been able to get the Dumpster Drobo’s (an old Drobo b800fs) throughput up to 30MB/s writing and 35MB/s reading, on large files (1GB+). But small file performance is still dog slow.

Even for its age, this thing shouldn’t be *that* bad. There must be a problem somewhere.

But at some point it’s a question of whether it’s worth spending the time to figure it out.

Show thread

Starting to think that my "dumpster find" Drobo may have been in that dumpster for a reason.

Too bad, as it *looks* pretty dang cool.

Definitely just remembered I have an early morning meeting that’s over an hour away tomorrow. That feels like an unnecessarily harsh way to start the week.

Good discussion on about @moxie Marlinspike's blog post re "web3" stuff. I wrote an overly-long comment that might be of interest to some:

tl;dr: I think Moxie's conclusions regarding web3/blockchain stuff are correct, but I disagree with the blanket assertion that end users can't/won't run "servers". Users absolutely will, *if* they have the right UX.

What most users don't want is a huge, loud rackmount "server". It doesn't have to be like that, though.

Ars Technica, not pulling any punches re the Google/Apple iMessage spat:

"After ruining Android messaging, Google says iMessage is too powerful"

Don't get me wrong, I'd be glad if iMessage was a cross-platform, open standard. It'd be nice if such a thing existed! But Google throwing shade at Apple for not cleaning up Google's own fucking mess in this area is… a bit rich.

Show thread

The doth protest too much, methinks:

It would have a bit more weight coming from a company that wouldn't absolutely do the same thing, given half a chance. And which has, through its long-running and ongoing incompetence, contributed more to the fragmentation of the messaging market than just about anyone.

Dear Google:
A Former Android User

Charles Stross' year-in-review / upcoming-year-prediction article is pretty good:

Particularly fond of the line: "Everybody loves unregulated derivatives markets until their imaginary wallet full of monkey jpegs gets stolen."

Unfortunately the rest of it isn't particularly funny.

Ideally, there would be an ISP that didn't meter bandwidth use between local customers on its own network (since it costs them nothing in IP transit). If that were the case I'd just get two accounts with the same ISP and problem solved.

But I'm not aware of any that work like that. They all seem to treat bandwidth across the street the same as across the planet.

Show thread

Question for experts and enthusiasts:

What's the cheapest way to get an uncapped (no transfer maximum) IP connection between two points? Assume in the same city, but hacks like WiFi or microwave dishes are out (no line-of-sight), and it's not close enough to physically run cable.

In the old days (1990s), you'd go to the phone co and get a "dry pair" between two points. Not dirt cheap, but not outrageously expensive. Many businesses used them, but it's all done via VPN now.

Good article on the dangers of layered abstractions hiding complexity until nobody understands it:

> [M]ost people would think that if we’ve built, for example, a space ship or a complex airplane in the past, we could build it again at any time. But no, if we weren’t building a particular plane uninterruptedly, then after just 50 years it is already easier to develop a new one from scratch rather than trying to revive old processes and documentation.

Yeah, sending an angry email is fun and all, but passive-aggressive faxes are really where it's at.

Wrote a blog entry with some additional thoughts about the 50 Year Computer ():

tl;dr: I think the technology probably exists to build it today, as long as quantum computing doesn't render all current numerical cryptography obsolete in the next few decades.

But it's a really fun thought experiment, and looking forward to discussing and writing some more about it in time.

Show thread

I may be coming in late to the discussion, but I just ran across "The computer built to last 50 years" [1] and "Decades of Fun: Computers Built to Last" [2].

A '50 Year Computer' is a great concept, although I disagree on some of the particular design ideas.

We are probably *close* to being able to build a very long-lived personal computer, absent any huge developments in cryptography. That's what dooms most old hardware today.


Well, I accomplished pretty much nothing today, making it an ideal in my book.

I probably should go out and shovel the driveway, but I'm not expecting the greater DC area to recover from this storm in 24 hours. So that seems like a tomorrow sort of problem.

A very fun "Things I Won't Work With": Dioxygen Difluoride (FOOF).

I believe some of the experiments mentioned (or very similar ones involving fluorine), are discussed in the truly excellent book "Ignition: An Informal History of Liquid Rocket Propellants". (also avail as ebook)

Weird to think that my prime music-buying years (before I had much else to spend money on) coincided with the high water mark of the industrialized music industry as a whole.

And that the industry today is taking in only about *half* of what it did at its peak. But yet, there’s no real shortage of . Funny, that.

It’s almost like they were being disingenuous during all those late 90s lawsuits.

Just in case anyone needed a bit of extra schadenfreude to cap off their year, here's a leaderboard of the biggest / / hacks:

Seems like an expensive way to collectively learn "huh, banking regulations exist for a reason!" but I guess some people need to learn by doing.

And yet the DeFi "industry" (lol) carries on. Oh, well. At least they're luring the more enterprising hackers away from extorting my phone company with DDoS attacks.

Show older
Mastodon @ SDF

"I appreciate SDF but it's a general-purpose server and the name doesn't make it obvious that it's about art." - Eugen Rochko