Thu 12 May 2005
Category : Commentary/versionTracker.txt
I've just updated Version Tracker and MacUpdate. I hadn't updated them for a while but they seem to be able to pick up the Postfix Enabler for Tiger releases all on their own. But they had the links wrong. So I had to go update them.
I saw a comment on Version Tracker just now about the password or serial number not working on the login screen.
This is puzzling because I never had a single e-mail about this. One would have thought that a person in this situation would have first contacted the author, having paid good money for something that didn't work.
In times like this, I think we need to go back to first principles - with the acceptance that people are not actually buying, say, Postfix Enabler. They have a problem they think Postfix Enabler can solve. So they're buying a solution to their problem.
Following this, I've made a refund to a person living in a dorm who did get his SMTP server to send mail, but he realised that he still had to go through the Smart Host, so there wasn't any point in using his own SMTP server. The software worked, but he didn't get what he was looking for - so I returned the money. That's sort of working according to the principle.
You know when Postfix Enabler works, it works beautifully. But there were a couple of cases where I couldn't tell why Postfix Enabler wouldn't work on those users' machines. It's quite difficult to debug these remotely. So many things can go wrong in the system (the network settings, the network itself, no available DNS server, a case-sensitive file system, a machine without a host name, manual modifications to the config files that Postfix Enabler is too stupid to work around, plus a zillion other things). But I think people appreciate it when I do give it a go anyway because it's really not worth the 10 bucks. So I'm actually relieved to be able to refund the money when things don't work out. There's a limit to what I can do a few thousand miles away.
So, I try my best to do the right thing but I am still disappointed with that comment on Version Tracker. I'll be glad to be shown where the login screen doesn't work and fix it. Until then, the evidence from all these downloads is that it does work. But it does leave a trace of fear, uncertainty and doubt, doesn't it?
Postfix Enabler for Tiger Release 1.1.2 is out
Category : Technology/PFEforT112Released.txt
I've released version 1.1.2 of Postfix Enabler for Tiger. It fixes SMTP Authentication via the SASLDB method, which broke in Tiger.
This version also contains an updated French translation from Michel Pansanel of Carpo.org. Thanks, Michel, I really appreciate it.
About PayPal and Automated Workflows
Category : Technology/PayPal.txt
It's been a week since we've had this system going which allows us to process credit card payments from almost anywhere around the world.
But using PayPal is probably the easiest way to go. Technically, it's relatively easy to set up and you get to test the system before going live. From the business end of things, it's easy on the pocket because there's no signing-on fee, no set-up fees, and no monthly fees. You do have to pay 6 cents off each dollar of the transaction to PayPal. And there's a sizeable exchange-rate conversion charge when you transfer the money to a local bank. But the pay-as-you-go approach makes it bearable.
One significant factor that has made life easier for me is PayPal's policy on refund. When you make a refund, PayPal also refunds its transaction charge. What I understand, from a traditional credit card arrangement, is that the merchant has to pay a fee to the credit card company for processing a refund. If this were so, every time I offer to make a refund (because I can't figure out why Postfix Enabler isn't working for a customer), I'm doubly punished by the system.
If you've ever tried to build a credit card processing system using the other (brain-dead) companies' systems, you'll understand what I mean when I say here's a breakthrough - on top of the fees after fees, you'll meet obstacles after obstacles just to get to test the other systems.
I think, then, that this could grow into something big. PayPal says that it has over 63 million accounts - which should make it a big enough ecosystem for people to trade with each other.
We've started off collecting donations from people happy to use our software. That gave us some time to think about what we need to do to make the transition towards actually selling things on the 'Net - which involves collecting money before we turn over the goods, in this case a serial number that will open our software.
The complications here are : 1) Timing issues. Sometimes, when PayPal is too busy, we get the notification that the transaction has completed very late, making the customer wait too long for his serial number. 2) E-mail issues. For any number of reasons, sometimes the e-mail with the serial number doesn't reach the customer (because he supplied an invalid e-mail address, or he has a picky mail server that bounced our message, etc...)
And there's a major technical issue concerning synchronising the transactions (keeping all the concurrent sessions apart), and avoiding deadlocks because they're all going thru a single serial number generator.
It's been a week and there's a lot to be learnt just from doing these - not only the technical issues but also the business issues - like how we can reduce the number of support calls - by either improving the user interface or simply improving the documentation?
So, in terms of the range of experience gained, I think doing a stint as a would-be on-line entrepreneur may be almost as good as enrolling for an MBA course. The Journey is The Reward ? ...