Comment by davidjgraph | Hacker News Reader

Comment by davidjgraph | Hacker News Reader

We developed a product specifically for Google Drive using their JavaScript client. We were hoping to make it commercial but had to abandon those plans due to 3 bugs that have been present since around the beginning of 2013.

Also, there's no open bug tracker. The developer relations people are very friendly, but when I ask where to report bugs I'm told "just tell me". So every 2 months I've reported the same 3 bugs in person and they get swallowed up and that's the last I hear about it until my next report.

The silly thing is that we were one of the showcased applications in Google I/O this year, for the Drive section. A number of Googlers came past and said "Wow, cool. Please, please commercialise this". The truth is, we can't and it's because of the bugs in Google's code (well, partly, the other problem is Drive has just been too unstable over the past 4 months, until there's 99.9% uptime for a year, we can't consider it).

Don't get me wrong, I love what Google are doing with their API stuff and the dev relations really go out of their way to help (specifically the Drive SDK team). But, as it stands, this isn't a platform I'm confident in building a critical business product on, currently.

EDIT : That last sentence is wrong. I'm confident to build a critical business product on it, I think it'll get there, I'm just not confident to charge money and have to support it, at this point in time.

The instability/unreliability of the Drive API can't be understated. I used to have almost my entire app rely heavily on gdocs, but the api is far too shaky. Even wrapping every API call in a try_5_times function (which would often take 2-3 requests before getting a non-500), server errors would get through. Critical components of the API would go down for hours -- once, the entire day -- sometimes without even a blip on the status page.

While the api is a little funny, it's clearly very powerful. Anyone criticizing the move to work with Google API's at all should keep that in mind. But the nonexistent service and unreliability of the API are hard to overcome.

I think you meant "can't be overstated" not "under stated" there.

Hah, yes. Thanks.

Can you link the bugs/stackoverflow you are having trouble with, please? We should be doing a better job of fixing them - especially if they are gating you from releasing a serious app. That is bad, sorry.

Hi Ali. [0][1][2], 1 is actually the dual clicking required on account selection using the JS client. They may seem subtle, but as we reduced our bug count these probably account for 75% of tickets.

TBH, I don't really want it to be that I whine in a very public place and the bugs I personally need fixing get priority.

What I'd really like to see come out of this is a public facing bug tracker for the Drive SDK that people can vote on and priority be given to the most voted for/most critical issues as viewed by the developers.

I can throw one onto github, as long as you don't mind something unofficial.

[0] http://stackoverflow.com/questions/13971272/is-there-a-way-t...

[1] http://stackoverflow.com/questions/13366254/is-it-possible-t...

[2] http://stackoverflow.com/questions/15664523/cant-get-google-...

Thanks for this. It's not whining, we take feedback extremely seriously. It is generous of you to provide it.

> until there's 99.9% uptime for a year

And three nines is still 45mn of downtime per month.

wha't s a "mn"? mili-newton?

He meant minutes. 45min.

That's usually acceptable. Annoying, no doubt, but those 45 minutes a month don't usually happen in 45-min blocks monthly. It's more like a few seconds every day, and a sparratic issue once or twice a year causing a few minutes.

SLA's commonly agree to 99.9% (and I've dealt with some picky clients), though 99.99% would be better, it's an acceptable amount. Servers break, people need to deal with it when they do, and we aren't all robots.

After reading through this thread it appears to me that all that is stopping you from launching is a couple of bugs related to multiple Google accounts use case.

I would call that an edge case and launch the rocket. Just drop a comment about it in the FAQ.

Unless, your product is strictly about multiple Google accouncts in which case you are shafted. :)

The part of the Google Drive market you can make money from with our application (it's B2B) is the Google Apps for Business accounts, that means Google Apps domains accounts (not personal ones). The majority of users logged into Google Apps Domains accounts (and we measured this) are also logged into their personal account on that same browser. This means more than 50% of our target market can't share documents through our UI, can't use iOS 6 and get a confusing experience logging in elsewhere.

That's 50% of users. All we needed was one user out of however many were on the domain to remark and the feedback was "sort out the details" and these were the details.

As also mentioned, the availability of Drive documents to our apps and the app integration working correctly was also a factor.

For what its worth I hope you manage to push through.

I think that Google is wasting huge ammounts of goodwil through its lame operation of GAB.

One does not simply not run a customer service.

Edit: To add a bit more substance to this comment. I kinda know wahat you are going through. I took it upon myself (after numerous have failed) to push through something that looked like a pretty nasty bug in Lotus Domino 8 server and client. Basically what I found out during research is 20% of IBM 852 (its internal datatype) was simply corrupt.

And boy for the longest time nobody responded. There were IBM SWAT Team Leaders guarding the sacred Process that would insist on telling me that there is no Bugs in their platform and that it is obviously my fault. It took six months, an extremely detailed testing procedure with test cases and invoking a lot of connections to get my report infront of a person that could actually understand what is going on.

Soon after that I received a hotfix. I had to repeat the whole procedure another two times. And by the third time I could get it done in three months. I basically made the blue elephant dance just for me.

So for what its worth my advice is to invoke every and all connections within Google not to just get your issues within an internal Google issue tracker. You need to find out (from the outside) who is the guy responsible for the general area that is causing you problems, then keep in touch with this person and after some time you will get your fix(es).

"I love what Google are doing" - but of course you do, it is called Stockholm Syndrome. Irony aside, i'm trully amazed of Google. See, the smallest misstep Apple makes, everybody jump to hate Apple. Google on the orher can do whatever they want and nada, everybody jump in to alleviate the evelasting, unconditional love, no matter what. Remarkable achievement, hat off, in todays world.

Don't patronise me. I've been developing commercially for 20 years now, I'm not exactly wet behind the ears, nor easily persuaded by anything but facts.

I've evaluated what they are doing on their technical merit. If you want me to give you a detailed explanation of why I prefer their system, it'd probably take about an hour, so it's a bit much as a HN comment.

I genuinely think their direction with the API I'm using is the way forward, I've been waiting for someone to have the balls to do it for some time.

Any anyway, I must have missed the outpouring of love Google received for closing Reader, for not having a native Google Linux kernel OS Drive client, and so on.

> I've evaluated what they are doing on their technical merit

That's your problem. Google is phenomenal on their technical merit. It's the softer, people skills (support, listening to bug reports, etc.) where they constantly shoot themselves in the foot.

> Google is phenomenal on their technical merit

They turn into Microsoft where brilliant engineers are bogged down by corrupt management :)

Funny that all the Google product managers I know personally are ex-Microsoft guys.

Does that mean that they brought Microsoft's problems to Google, or that Microsoft is having problems because they left? ;-)

Google is phenomenal on their technical merit.

It is precisely technical merit where they fell down and that's why he decided against the development.

I score Apple and Microsoft pretty low on these two measures also. Support MS are OK to good, but listening to bug reports seems to be hard to do at scale. Can anyone provide any examples of people doing it really well?

Linux kernel.

<<Google on the orher can do whatever they want The funny thing is: That was Apple, about 3 years ago.

...and it was Microsoft about... never. I remember people complaining a about Microsoft around MS-DOS 3.1.

It's snarky, flame-ware inducing comments like this that are bringing down the the quality of conversation here.

Google has flaws like all companies and individuals. I'd be careful to share sensitive information with them. That said, they have a long track record of contributing back to the community (e.g. release stuff as open source like chromium); generally support open and royalty free standards (WebM, SPDY); and generally do nice stuff like Summer of Code.

If they'd turn evil over night or just stop existing, we'd still have a big net positive left. Not something that can be said for some competitors.

This is a joke ? WebM and SPDY aren't some altruistic gifts to the world.

WebM is just an pointless nuisance. The world already has a open, video standard that is widely supported, technically superior and stewarded by a wide array of competing companies. And it's only royalty free because nobody can be bothered to sue them yet. It definitely is patent encumbered (see MPEG-LA's recent 'gift' of patents to VP8).

SPDY was another technology that was developed by Google unilaterally instead of working through the existing IETF process. If Google didn't own Chrome it simply would have never taken of.

It's wrong to place SPDY in the same class as WebM: it actually provides a significant, meaningful benefit versus the status quo. The performance benefits alone are worth deploying it for most sites and it's having a big, positive influence on the standards process as people learn from real-world deployment experience and start discussing how it makes web development better.

To be fair, I think there was a mix of altruism and self-interest with both technologies. Without WebM being a credible alternative, I think we'd have seen H264 licensing tighten up considerably.

And the fact that SPDY was developed unilaterally doesn't take away from Google's generosity in making it available to the world. It's generally well-regarded, and Google haven't attempted to limit it's use by other browsers in any way - much the opposite. Can you imagine Microsoft doing the same when they were the dominant browser?

Again, I'm not saying they introduced either of these technologies from pure altruism. But I am saying I'm glad they exist, and Google's approach to IP is a welcome breath of fresh air (so far).

Heh-heh, this comment of mine has the most interesting history. First it went straight up to +12, and now it has come down to -4. So it has my personal record of minimum 16 downvotes. There were some bumps when it went up, so it is certainly more than 16, only i do not know how many.

Anybody has collected more down's for one comment, or is this some kind of record?

At risk of doing the same, please stop mucking up this thread with irrelevant/flame-bait comments.

You're wrong in every detail. Both Google and Apple have a lot of haters and a lot of unconditional defenders. This is, of course, because they both do some shitty things and some awesome things. But good luck with your effort to cram the world into simple black/white cubbyholes!

Reply on news.ycombinator.com