Made some good progress on the adventure game. I have a pretty silly split of files though...
game.py -> main file
actor.py, inventory.py, menu.py, scene.py, uihandler.py (decided against actions because I also do stuff like change the cursor etc. there)

Would you keep the split (essentially 1-2 classes per file, menu.py has a Menu and a MenuItem class for example)? I'm thinking I'll keep it this way and put them all in a package like "adventurelib" or something.

I rarely do event driven programming so the entire architecture and layout seems very kludgy to me but I'll do a major revamp once I understand how testing really works in this setting

Shoe Lace,
Your angle here with these low end clients should be to try and get them to host with you. You find the cheapest way you can get their hosting done and then charge them a $30-50/mon fee depending on how much you think you can get. Sell them on the "yeah, but if you host with the $5 plan you're going to end up paying for support."

At some point you'll have quite a bit of recurring cash rolling in and can take a break from development whenever you want.

The cheaper clients either have their own host or I do host for them but I don't markup the hosting price. I only bill when support/etc. is needed.

There's a lot of sites that I haven't touched in a long time because a properly setup host requires no support or maintenance. It's not uncommon to have 500+ day system up times even while pushing new builds to the site and upgrading versions of various services.

If they want a new e-mail address or something setup then I would just bill them on demand.

And even with 20 reoccurring clients at $30/month that's only $600/month. It's not bad and certainly better than nothing but it's not close enough to stop working.

I don't understand... it's essentially $600 more money for you for very little additional work, and your argument for not doing it is that it's not enough to stop working?

As GM points out, it's free money. I do like that idea of being your own heroku (you have the skillset to do it) and just undercutting them a bit.

I think you still don't value your skill set enough! Web development is super in demand, and people are willing to pay tons of money to smart people they can trust. I used to be in the boat that $60/hour (not even a year ago) was a lot of money, and now if I'm doing consulting work, I'd probably scoff at less than $100/hour unless it was a longer term contract.

It makes marginal differences in the cost of hosting seem less important, since I'm not trying to deal with clients that are cheap. Don't be afraid to ask for more, the worst they can do is say no or try to negotiate (in which case you'll prolly end up with a higher hourly).

Shoe,

If you haven't read this you might get something out of it: http://www.kalzumeus.com/2014/04/03/fantasy-tarsnap/

Basically the classic tale of the smart tech guy who misses tons of value because he doesn't understand how businesses view value.

It's no additional work other than processing the bill every month because I already do it. I just usually factor in the cost up front in the bill for the "work" it might involve to set it up once. Like I might just say to myself "ok, I need to setup a VPS for this guy, that'll be $300 I'm adding to his bill" and then he still pays the hosting bill every month on top. For non-programming heavy jobs where shared hosting works and I throw up a simple WP site then I just put all of the hosting costs under their details.

GM,

Read most of the article. I dunno. His new pricing table seems more complicated to me.

Look at EC2, it's pretty much designed to let you know how much it'll cost as long as you know what you're doing and have things planned out. That makes sense to me.

Telling me I have unlimited storage but only 500 GB of media for $100/month doesn't even make sense. Also what if I decided to zip my media, now it's a zip file. Does a zip file containing media count as media?

I like his old payment style much better because if he's focusing b2b then someone should have the skills the determine what's actually going to be backed up and how big it will be. If they don't then someone needs to be fired and they have a lot bigger problems than backups on their hands.

It was a pretty good article overall tho, I read 75% of it.

Just a guess, but I bet you could keep your price the same, and then advertise that you do cheap, fully managed hosting for a price you won't find anywhere else: $30/mo. "I've never had any of my servers go down, but if does, just call me and I'll get on it."

Again a guess, but you probably think that sounds underhanded or something, when in fact it actually is a good deal for a nontechnical person. Sure, it costs you $0 and you are charging $30 and for the most part not spending time on it, but it's not like they have alternative that are giving them a better deal. It's completely fair from their POV, and pretty cheap.

You know what heroku charges for the same service? $1000/mo. They have guys who wear pagers, but basically the service is: If some **** happens we'll help you at within X minutes. I guarantee most of the clients spending that are just paying an insurance bill every month.

I mean, I think this just confirms that you have a strong techie worldview and aren't looking outside of it. You are Colin's ideal current audience. Patrick's whole point is that catering to that audience rather than a business minded one will leave massive amounts of money on the table. *And* it's not what the business guys want anyway.

You seem to be running your business with the mindset of "what would I want? what do I think is fair?" instead of "What do these other, totally different minded people want? What would they think is fair?

Yeah that's a fair assessment, I won't contest that. I see it like this:

Let's say I offer managed hosting for $30/month and I'm really paying $8/month, so now I'm profiting $22/month but what value do I provide in this situation?

If it's a shared hosting provider and the site dies then there's nothing I can do about it. I don't run the servers. The best I can possibly do is parrot info from the hosting company to the business owner and hope the shared provider fixes the problem asap.

With a VPS it's pretty similar honestly. I've never had ubuntu just lockup or postgres crash which involved manual intervention. It's rock solid, especially if you're not really pushing the limits of the hardware.

peace of mind. transfer of responsibility. basically, they don't want to think about it and the value you provide is telling them they don't have to. You aren't cheating them either, because if something does go wrong you're going to fix it.

a) parroting info they can't understand into clear, bottom-line language *is* a service.

b) if **** really goes down, you have the skills to redeploy them to another host fairly quickly. that's a very real insurance policy.

i can hear you saying, "but they might be down a *day* or more in that case." well, that case isn't that likely, and that's pretty damn reasonable for $30/mo catastrophe insurance. you remember that 1000/mo heroku premium support i mentioned? guess what's not included? DoS attack protection. When I asked them about that, they said they'd do their best to help me figure it out if it happened, but it's not actually included.


this is MORE reason to do it, not less.
Shoe's mindset: "well i'm charging $30/mo for doing nothing, the servers are so solid. that's not fair... what if they find out?"
Their mindset: "this dude is awesome, he charges basically nothing, $30/mo and i've never had a single problem."

Theres a lot of value to non-tech people in just knowing there is a guy you're familiar with that you can call. People really hate filing tickets through their hosting company especially when the site isn't working.

That's the thing though, when things go wrong I don't fix it. It gets fixed by the provider and 99.99% of the time it gets fixed on its own in minutes/hours. The business owner doesn't care who fixes it or what the problem was. He only cares that his site is accessible. It's straight up lying to the business owner if I tell them I'm managing their server. That implies I have direct control over everything and it's on me to fix the problem.

It's rarely a case where the downtime is extended enough to warrant a ticket. I can only think back to 1 time where I opened a ticket and it was was some obscure as balls incompatibility with querystrings, an htaccess rewrite rule and a cron job.

What usually happens is maybe the site starts to 500/404 and I get a call from the business owner. He tells me the site is down. I do very basic checks over the phone in real time (try to load the web site, check the hosting provider's status page, is it down for everyone, etc..) and then either:

1. I usually offer to call the provider and get answers in exchange for $.

2. I'll find that the downtime is known by the provider and wheels are in motion to fix it by finding it out from their status page.

It really is like insurance. Useless 99.999999999999% of the time. I just don't charge for insurance. I instead charge more up front and just add it to my bill and then deal with problems as they come up for either billable $ or spend a minute or 2 and give free value to the client by looking things up for them.

you dont have to lie, and it's not lying. they'd be paying you to email with support at xxxHost or whatever because they don't know how the **** to talk to some tech support guy on email. think of yourself as a translator, or a point guy. those are real services.

obv run your business however you like, it was just an idea i thought might be helpful.

EDIT: btw you can present it totally straight up to them. "After you launch, you can take over hosting and talk to support at xxxHost yourself, or pay me $30/mo and I'll handle all communication for you and call you and explain in simple terms any problems, etc." I bet you'd be surprised how many people take you up on it.

It sounds like you're already giving them the tech support. Think of it as just billing them on retainer rather than a strict hourly rate.

People would take me up on it, no question about it.

I sort of do get the value from them, it's just done up front. If I adjust the price by $300 then it's like I'm getting 10 months of $30/month.

I am losing money in the long term tho, it just sucks to think about because for the amount to be non-trivial you need so many clients and they need to keep their site going for years. Profiting $20ish per client means you need 50 clients to just hit 1k/month and at 1 new client per month that's 4+ years.

Shoe -vs- The World.

The World is wrong because offering tech support @ $30 / month is ******ed. It only takes one needy client to decimate your hourly wage.

Shoe is wrong because obviously this is a good money making machine. I've worked with lots of non-tech people. Trust me, you have no idea how much the comfort of knowing there is a human on the other end is worth to them. When you have people that are utterly afraid of their computers to serve, which is likely if they aren't paying much for your services, you will be the world's best programmer. They need humans, you don't. Don't ever forget this.

How can you seriously write that?

You realize I am charging people right, but instead of getting $30/month I get 10 months of that up front. I did state that multiple times now, I'm not sure how much more clear I can make it. I'm willing to sacrifice literally 2 minutes of my time on something that happens like 0.000001% of the time for the money up front.

I've never once been in a situation where I've gotten "decimated" over it. Also that statement is implying that I don't charge people for other non-server issues but that's wrong because I do. I'm not going to work for free and I always limit revisions/etc. and have contract clauses that make people aware that extra work will cost more money.

It's also more like forums vs me because I simply stated that you cannot say your time is worth money until ALL of your time could be billed. Somehow that ended up with a bunch of people questioning my billing structure.

Of course your time is worth money regardless of how much of your time you bill for. You are providing a service at a price that doesn't scale and are one heartbleed type event away from taking a big hit on it. If you can support your customers issues in less than an hour a month you're in a good situation. Not as good a situation as if you charged a more reasonable rate for support though.

do you have a line item that says $300 -- 10 months server support?
do people think your overall package price is a little high until you point that line item out?

unless both of those things true, then you aren't charging 10 months up front. you just have some idea of that in your head, and in actuality the client would have paid your entire current price PLUS $30/mo (or more, who knows) and still be totally happy and think your prices are reasonable.

btw you're right that none of us know **** about your business and are just armchair pontificating.... so i'll qualify the above again that it's just kind of guess based on the way you write about these things. and it's not "questioning your billing structure" -- it's like hey dude, maybe you could be making a nice chunk of extra money for almost no work.

so after watching gary bernhardt's talk on the birth and death of js, it got me thinking about asm.js, which i haven't really used. anyone have experience with it?

would you be able to, say, write a fast hand evaluator in C, compile it to asm, and then run stuff like pokerstove calcs directly in the browser at ~50% native speed?

created a hand evaluator using objects (i know, facepalm, byebye performance, just did it this way as it just felt natural to code the cards like this at the time derp).
what would be the benefits of this way compared to twoplustwo hh eval which uses bits and lookup tables? anyone know how much memory the eval takes up? (im guessing its negligible).

depending what language you're using, objects can be fine if your design is right. a while ago i rewrote the 2p2 hand evaluator in Java, and had a serialized file for the LUT. iirc it was ~230MB or so, so not bad at all. if you are ok with having that kind of memory in RAM (these days it's kind of nothing) then the hand eval code becomes really simple if you take the generated LUT as a given.

not sure if this approach would work for a browser/asm based evaluator -- anyone know what the limits of localstorage on chrome/ff are?

I just slap it onto the final bill. There's not a trace of it on the invoice.

It's always possible to make more money. I should probably go after the extra $20/month because money is money in the end.

Not offhand but Opera 12.x warns me about local storage if the amount is over 5mb. It prompts me asking me if I want to allow the site to use more.