4 ms·
I clicked through and thought you were being too harsh. I mean, it can be fun to make toy implementations of things as an exercise. Doing an SSH server in PHP w
by wpietri 13y ago
I clicked through and thought you were being too harsh. I mean, it can be fun to make toy implementations of things as an exercise. Doing an SSH server in PHP would be entertaining if you liked PHP. You'd learn something.
And then I read that his hacked-together-in-3-days ssh server was for use in production. In a hosting service.
Wow. Just wow.
- dspillett 13y ago> And then I read that his hacked-together-in-3-days ssh server was for use in production. In a hosting service. That sounds like a brilliant technical guy, capable of running with a daft idea to completion (unlike me, with my collection of at-best-half-built personal projects), who should have some layers of protection between him and Real World Production...
- outworlder 13y agoThat reminds me: "Now, people who Get Things Done but are not Smart will do stupid things, seemingly without thinking about them, and somebody else will have to come clean up their mess later. " http://www.joelonsoftware.com/articles/fog0000000073.html http://www.joelonsoftware.com/articles/fog0000000073.html
- eldavido 13y agoI think it's different. Inexperienced (young) programmers don't know what's been tried, and what's available. I've been dealing with this a lot at work recently, where we ended up doing poor reimplementations of off-the-shelf stuff due to a mix of ignorance, and honestly, a bit of hubris. It's a good attitude to have in academia/non-production critical work, but the GP is right, production demands a more conservative approach, especially when money/safety is at stake. The type of ignorance Joel writes about (see parent) is different, it's more like "slavishly following design patterns" and "writing copy-and-pasted, improperly factored code". Both are ignorance, but the first is better called "NIH", whereas Joel's is, "writing bad code".
- chii 13y agowell, is it better to have a hacked MVP released in production, or spend forever making it and never actually releasing it and then missing the window?
- jbrechtel 13y agoDepends on what the product itself is and from which perspective you're asking the question. At any rate, those are almost never the only two choices...
- mrcwinn 13y agoI can tell you absolutely, without question, that when it comes to security and people's funds, there is nothing courageous about a hacked MVP in production. There's a difference between someone's to-do app one weekend, and this case. If you are handling people's money directly or indirectly, you need to care about that and take it seriously. Or don't ship.
- eli 13y agoWell I would argue that the "M" in MVP would necessitate never losing anyone's money. If I were going to create a trading platform, I'd probably start with one that only accepted Play Money.
- wadetandy 13y agoHe did. It was called Bitcoin
- eli 13y agoCute, but they also traded and held USD.
- wpietri 13y agoPutting on my Lean Startup hat for a sec, I would even say that the M could allow for losing money. If it's early on and your customers are all in the 2 1/2% of innovators, they will put up with a little of that. Certainly if you make them whole, but probably even without. That said, "flawless accounting" would be very high up on my feature list. I think the failure here isn't launching without perfection; it's operating at scale without perfection.
- MrBuddyCasino 13y agoOmg, I nearly missed that: What did I create a ssh server for? The same thing I created a DNS server for fun and for KalyHost. I'm not sure "disbelief" is sufficient to describe how I feel right now.
- ansible 13y agoYeah, I'm still trying to figure all that out myself. There was some allusion to needing some kind of database backend for the SSH server, but there are multiple solutions for that now (like LDAP). I'd love to have this guy work for me in a junior role (because he can really crank out the code), but all his work would need to be reviewed, and I wouldn't want him to be making architectural decisions on his own.
- danielweber 13y agoThe part of "running a SSH server that you wrote in PHP in production" that is scary is not the "in PHP" part. It's the "that you wrote" part. No matter what language you write it in, you are going to mess something up. The OpenSSH guys have messed up working a lot smarter and more diligently and with more time than you have.
- ultramancool 13y agoThis is so very true, OpenSSH is probably one of the most secure pieces of software around. It has extremely high value to attackers, yet has had extremely few remote security holes in its lifetime. They've invested years and many talented people in developing such a piece of software. If you want to write your own ssh server in php, you should probably consider your motivation and how you can re-use their code or operate through it instead if your purpose if anything other than experimentation.
- MrBuddyCasino 13y agoYeah, thats what I thought too, until I read this little gem, suggesting that OpenSSL was written by monkeys: http://www.peereboom.us/assl/assl/html/openssl.html http://www.peereboom.us/assl/assl/html/openssl.html Its kinda hard to disagree with that conclusion.
- deciplex 13y agoOpenSSH is not OpenSSL.
- MrBuddyCasino 13y agoDang, guess that joke's on me. Wouldn't have realized it, thx for the correction.
- danielweber 13y agoYou shouldn't be using OpenSSL unless you are an expert in crypto and software development. It's not easy to use and it shouldn't be. I can't say those difficulties he had in using the library were put there on purpose to keep people like him out, but it seems to be a good effect here.