Litany against NAT
apologies to Frank Herbert
I must not NAT.
NAT is the net-killer.
NAT is the little-death that brings total obliteration.
I will face my NAT addiction.
I will permit it to pass over me and through me.
And when it has gone past I will turn the inner eye to see its path.
Where the NAT has gone there will be a proper firewall.
Only IPv6 will remain.
I have my phone set to auto-deny any number that is not in my contact list to combat the ever increasing amount of spam I've been getting. Unfortunately there is one insurance scammer that has taken to leaving voicemails. Multiple VMs a day for the last two weeks.
I contacted my provider about disabling my VM box altogether. They do not support this. You can, however, redirect missed calls to another number instead.
I now send missed calls to: https://github.com/pjf/rickastley
Eat Rickroll, spammers!
If people want an open network, stop shit-talking open things
I hear tech folks saying that RSS is dead, and people on social media saying that blogs are dead.
The problem with this is twofold:
- RSS is doing just fine, and not even remotely dead
- Blogs are doing just fine, and are not even remotely dead, either
There's this strange sort of defeatism around. "Well, we lost, so what can you do". Um, not undermine things that still work and are actively being used? Support them as well?
A comment I just made on IRC in regards to Linux distro maintainers disabling #IPv6 listeners in their packages:
RFC 8200 (STD 86) has been out for 2.5 years now. Disabling IPv6 is also a violation of RFC 6540 (BCP 177).
Never not funny. Kermit and Fozzie, in existential dialog.
The fact that the guy who signed off on this patch uses the domain trash.net is exceedingly appropriate.
MEMESat-1: A Meme-Beaming Cubesat Currently In Development
Today it was just npm, but a few months ago Cloudflare managed to disrupt access to most of their customers (which is a significant portion of sites and services on the Internet) due to a bug.
Bugs happen, that's not a problem. The problem is one bug at one company caused that level of disruption.
The old truism of "the Internet routes around damaged networks" doesn't hold anymore. That needs fixing.
"We don't need to support #IPv6. There will always be IPv4."
"We don't need to support UEFI, there will always be a CSM (BIOS emulation)."
"We don't need to move to nftables, there's an iptables shim."
I get it, change is hard, but how else are we gonna get rid of the cruft?
Also, I realize this is antithetical to the "reinvent the wheel" problem (SQL is proven tech, why do we need noSQL?)
How do we move forward in stable, principled ways?
my response to "Did the Early Internet Activists Blow It?"
My take is that his analysis and worldview are hindered by presupposing the existence of large government institutions and corporations, and of the basic architectures of the Internet and Web. His view is as an internet lawyer who showed up in 1990, so he has little to say about the early architectural decisions that engendered the surveillance capitalist hell we have today. He maintains that we need more "leaders and policymakers and journalists and thought leaders" to deal with the nexus of problems we face today. And so he fails to challenge the mental models and structures that created this mess, instead preferring to stay in his comfort zone as a career lawyer.
While there were real legal hurdles to tackle in those days, namely the cryptography export restrictions by the US, I believe we would have been better served if the architects kept working to achieve their mission of internet freedom of expression. They were satisfied with good enough and perhaps feared researching in the same directions as the cypherpunks. More likely that their mental models presupposed the same power structures and sought merely to work within those whilst receiving research grants from the USG.
The author recognizes the terror that is Google and Facebook and their cozy relationships with government powers but again wants to address the problem with public policy. In my view, the internet architecture should avoid centralization of key services because of the perverse incentives and damage to social freedoms that those system structures create. The author says they failed to predict these dominant institutions, but I am not convinced since the system architecture requires them to scale and that would mean the internet founders hoped the internet would never get big and important. These were unsolved problems in those days, but my opinion is that we should've researched harder in that direction than we did.
A cyber-architecture that favors large socio-technical hierarchies as we have today is (now obviously) going to create more opportunities for oppression than any optimistic view would hope for. We know now that we have to "lock the web open" because the powers of public policy and the state are greater than the current internet's ability to oppose them. We should be thinking more about federated, peer-to-peer, and other mixed decentralized models such as the fediverse for inspiration, not the rigid hierarchies and collateral damage of the Wikimedia Foundation and Wikipedia editors. While some detractors say that fedi in-fighting is tiresome and broken, I believe this is precisely how it is supposed to function and is more fitting to human behavior than large centralized services like Facebook and its army of AI moderators. Further is that the kinds of cooperative agreements and policies enforced by fedi instances are more appropriate and effective than any large sweeping countermeasures that a government or corporation could provide.
When 30-odd of us packed into a room at the Decentralized Web Summit 2018 with Cory Doctorow, he asked us: how do we keep the web from swinging between decentralization and centralization, or is that inevitable? I argued that we need blueprints for organizations and architectures that are known to be resilient to centralization creep, and noted that we tend to recreate systems with which we are already familiar (Conway's Law). So my conclusion is that we must re-architect the cyber-power-economic model to destroy the incentives that favor oppressive regimes so that the incumbent model will simply eat itself into irrelevance.
A crash course in the textmode spreadsheet sc(1) http://blog.startaylor.net/2016/08/29/sc/
GitHub has a /44 of #IPv6 space: https://whois.arin.net/rest/net/NET6-2620-112-3000-1/
They're even announcing it: https://bgp.he.net/AS36459#_prefixes6
But no AAAA records.
People have been asking for this for at least eight years. That's longer than Docker has existed.
What. Is. The. Damn. Hold. Up ???