Agree with you on the Google not caring about the product at all. I am so dependent on Google Voice right now to the extent that I don't even know my real number. But Google's approach to Voice is scaring me more with every passing month and year. I even switched to Sprint because they have carrier integration with Voice but even that is a half-hearted integration:
- Text messages arrive after random delays.
- My outgoing number switches to the Sprint number if I am on call roaming.
- When I am logged in to gmail on my computer, it is always unpredictable where it is going to route my incoming call - to my phone or to gmail.
- Some services don't consider the Google Voice number to be a valid number (for e.g. I had problems recently trying to get FiOS to recognize my Google Voice number)
Yeah for scientific computing specifically i was more expecting to see concerns like ensuring intermediate working state is persisted to a durable medium and facilitating restarting from a working state "dump" instead of from the top each time.
Around the end of summer there was a post on just this topic. I can't find it now but it had really solid practices in this area. If anyone remembers it / finds it I'd love to be able to read it again.
You may be talking about a post by S. M. Ali Eslami called Patterns for Research in Machine Learning. There was some discussion here as well.
The patterns he pointed out were:
1. Use version control.
2. Separate code from data.
3. Separate input data, working data and output data.
4. Modify input data with care.
5. Save everything to disk frequently.
6. Separate options from parameters.
7. Do not use global variables.
8. Record the options used to generate each run of the algorithm.
9. Make it easy to sweep options.
10. Make it easy to execute only portions of the code.
11. Use checkpointing.
12. Write demos and tests.
His post is mistaken. His account isn't banned. I banned it temporarily while waiting for a reply to my email, then unbanned it. (I often do this when I ask people to stop doing something, because I've found that some people keep resubmitting otherwise.) I don't remember exactly how much later that was, but usually it's less than a day; if it were longer I'd forget. If he'd replied to my email I would have unbanned him immediately.
I told him this when I came across that blog post a while ago, but he seems to prefer his more colorful version of the story.
I think the problem here is that the only type of banning you have is hellbanning and slowbanning. It would have been clearer if you had just “normal-banned” jcs, so he got a message saying “you have been banned and can’t post anything”. That would have avoided the misunderstanding to some extent.
Hellbanning should be reserved for users you want to prevent from noticing that they’re banned, so they won’t create a new account. That would be users who you think hold no hope of salvation, not those who you are just giving a warning to.
Even better in this particular case would be yet another type of ban that informs the user “you have been temporarily suspended from posting; check your email for the reason”. That would let the user know that the ban is temporary so they don’t get upset that they were apparently permanently banned for one mistake.
Yes, it would be nice feature to have two kinds of banning. But there are so many other things on my todo list, and though this particular glitch attracted lots of attention, it happens vanishingly rarely and doesn't affect the site much. So there is a lot more payoff for users if I focus on e.g. ways to improve comment threads.
As an aside, you should consider disabling the automatic IP banning, which is clearly overzealous. The places I read HN from, including the office and my home LAN, have static IPs, and approximately every 14 days I find that my IP is suddenly banned, and I have to resort to a proxy server to use HN.
When this happens I end up shooting you a personal email, to which you have replied, dismissively, only once. Either you are reading my emails or there is some kind of expiry on the bans, because I find I am usually unbanned within a few days to a week.
Oh, and there is nothing malicious happening on my LAN that warrants this kind of banning. I'm a casual HN reader/poster.
I got the same treatment he did -- and have on accounts I have made since -- and we exchanged emails during development. We both went through each others' posts and couldn't find any rational reason for the treatment (slow banning and hell banning, in both of our cases).
Lack of transparency is a pretty big deal on this forum, and once you've crossed Paul -- without knowing how -- it's scorched earth on any account you make. That's why Josh developed Lobsters, and why I'm user #2, and why I'm glad to see it's gaining traction as a community.
This site is more threatening to opinion and discourse than any on the public Internet, and that isn't hyperbole. The irony is that this is a hacker community, and most hackers would be appalled at slamming the door and creating a curated garden of ideas. I've posted as jsprink_banned and jsprinkles on the topic, if you're interested; won't spam this thread with it.
Graham allows the dumbest people on the Internet to come onto this site to berate startups he all but begged people to create, but hellbanned a guy who pooped out a better version of HN in his spare time, for nothing. For fuck's sake.
My first account was hellbanned for no obvious reason. I had a bunch of karma, hadn't been downvoted, hadn't trolled. I asked pg for an explanation/reconsideration...nope!
One of the dirty secrets of hackernews is that you can be hellbanned at any moment without rhyme or reason, and you'll likely never know why. I assume it serves some purpose OTHER than driving away helpful contributors, but I'm not sure what that might be.
I think it's fair to say that hellbanning is done to improve the site, rather than for the nefarious purposes you're trying to ascribe for it. Clearly, there are problems with it's application and transparency, but that's not the same as suggesting there's some plot here.
No, it's not fair to say that hellbanning - as implemented on HN - is done to improve the site. It's arbitrary, capricious, and hugely non-transparent.
No, I'm not suggesting a plot, but there's no natural law that says tech news sites must have a hairtrigger hellban (nor, as far as I'm aware, does any other site have anything remotely like HN's policy). Concious or not, it's a policy that HN has adopted, and it's a terrible policy.
To what end? Paul doesn't want me around and has made that passive aggressively clear, so I'm not sure what good contacting him would do. I discussed on jsprink_banned how not spending immeasurable amounts of time in HN threads has had a positive impact on my life anyway.
Wouldn't it have been easier to just register a new account on hn, or just treat it r/o and don't bother registering?
Unless, of course, PG has written some software that identifies your browser across accounts (which isn't terribly hard,) but as you say there's no transparency here, but then it is his forum so I don't expect any kind of transparency. He can do what he likes and ban whomever he feels like.
I commend the effort, so far the content looks like it is a clone of hn pretty much, I recognized many of the titles, but I think it's great you've created your own version.
The section in the penal code should have read "deliberately letting your religious feelings get hurt and attempting malicious acts intended to outrage the religious sentiments of any class or community." instead of "deliberately hurting religious feelings and attempting malicious acts intended to outrage the religious sentiments of any class or community."