3 ms·

Deprecating API's with repeated warnings and a long wind down period is perfectly acceptable. You were able to use the old Datastore & Blobstore for what, years

by recuter 13y ago

Deprecating API's with repeated warnings and a long wind down period is perfectly acceptable. You were able to use the old Datastore & Blobstore for what, years? The $500/mo support plan is comparable to Amazon EC2. The email quotas and free tier in general were very generous, I don't understand the complaint that you can "only" send 100 e-mails for free - Amazon SES, Sendgrid, and all the others have similar limits in place. Emails cost fractions of pennies to send and its a simple API or SMTP, you can switch to any of the other providers. Bing to find one that gives you 12k transactional emails for free. :0) I agree the pricing changes shakeout came as an unpleasant surprise, but there's no free lunch. At least its a profitable/viable product now and won't be killed like Reader, I'm actually happy for that long term.

jwheeler79 13y ago
Another major problem I have with the Datastore deprecation is the precedent it set. Datastore was a GA feature AFAIK. So, there's no real distinction between GA and experimental now. If something doesn't work out for Google, experimental or not, they're going to can it--that's the message they sent. They should've grandfathered people over and just bit the bullet. Maybe come up with a rule like: if you don't write to the old Datastore for 90 days, you never get it again. With that being said, long grace periods don't matter when you've got a terabyte of data in that thing--For me and a lot of developers, it's going to be risky business to migrate from Datastore to HRD, and I dread it. I'm not paying them to dread their infrastructural issues. No. That's the opposite of what I'm paying them for. The problem with at first having practically unlimited e-mail and then the new restriction is it limits the type of apps you can build without their permission. I understand the need to protect against misuse, but you'd think they would've gotten it right from day one since they manage the largest e-mail service on the planet. I don't know if you were there during the early days or not, but one of their strengths was support was very accessible. You could write about an issue on the newsgroup and an engineer would contact you personally, go into your app, unstuck your indexes or whatever. Going from that level of service to $500/mo once you're hooked stinks. It's the old bait and switch all the way.

by recuter 13y ago

jwheeler79 13y ago

Another major problem I have with the Datastore deprecation is the precedent it set. Datastore was a GA feature AFAIK. So, there's no real distinction between GA and experimental now. If something doesn't work out for Google, experimental or not, they're going to can it--that's the message they sent. They should've grandfathered people over and just bit the bullet. Maybe come up with a rule like: if you don't write to the old Datastore for 90 days, you never get it again. With that being said, long grace periods don't matter when you've got a terabyte of data in that thing--For me and a lot of developers, it's going to be risky business to migrate from Datastore to HRD, and I dread it. I'm not paying them to dread their infrastructural issues. No. That's the opposite of what I'm paying them for. The problem with at first having practically unlimited e-mail and then the new restriction is it limits the type of apps you can build without their permission. I understand the need to protect against misuse, but you'd think they would've gotten it right from day one since they manage the largest e-mail service on the planet. I don't know if you were there during the early days or not, but one of their strengths was support was very accessible. You could write about an issue on the newsgroup and an engineer would contact you personally, go into your app, unstuck your indexes or whatever. Going from that level of service to $500/mo once you're hooked stinks. It's the old bait and switch all the way.