Hacker Newsnew | comments | ask | jobs | submit | kstrauser's commentslogin

Exactly this. I wrote a blog (at https://www.crittercism.com/blog/scaling-with-eventual-consi...) about why my company designs for eventually consistent stores from the beginning. There are a lot of use cases - and the number is growing as companies deploy large, scaleable systems - where a consistent store just isn't practical or even needed.

You can't just use an eventually consistent store as an ACID system and expect it to work, and it's almost always a Really Bad Idea to try to implement ACID on top of EC. Understand how your database works and design to its strengths instead of trying to pretend that its weaknesses don't exist.

reply


Why is that inappropriate, out of curiosity?

-----


I got tired of hearing friends saying that they'd rewritten their frontend in [trendy language of choice] because their backend was slow.

Bleem is a placeholder for these conversations. If replacing everything in an infinitely fast, no-resource-using language won't fix your performance issues, then maybe you need to take another look at your architecture.

Example application:

"Our API servers that make 1,000 concurrent requests to MySQL are slow. We're thinking of replacing them." "Would Bleem fix that?" "I guess not."

-----


So, am I (as #3000-something in the beta queue when I signed up for it) any nearer to having an account I can start testing? :-)

-----

mikegriff 22 days ago | link

Send me a mail, I have a couple of invites available.

-----


May I have one? I'd email you directly but can't find your address here.

-----

bitcrusher 29 days ago | link

Sure, my email address is in my profile.

-----


On OS X, I use DEVONthink Pro Office with a ScanSnap iX500 scanner, and does pretty much what you're describing. OCR is really good.

While it manages the directory structure, all documents are stored "bare" in its filesystem folder. That is, you can create a Foo.pdf in DEVONthink and then access it from ~/Documents/whatever/PDFS/Foo.pdf (or however it names the directories).

Machine learning classifies incoming documents and can suggest relationships between them.

I'm not related to them, I promise. :-) I'd bought DT a while back but hadn't used it much until I got Doo as part of a software bundle. I played with it a while, realized I already had DT which was much more powerful, and started using it more.

-----

rsanders 68 days ago | link

DEVONthink Pro Office doesn't have the beauty of Doo or a name that's even remotely decent from a marketing point of view, but it is a workhorse of a doc processing engine. And it even syncs between multiple computers and Dropbox for your own private cloud.

The only thing I'm missing now is an Android "client" for it.

-----


Nope. He said the average employment length was 75 days, so that 70K was total. Also, they're based in Toronto so that's CAN $.

75 * 6 devs = 450 dev-days, or 1.23 dev-years. 70K/1.23 = 56.9K. According to Google, that's $51K USD per year.

Yeah, I can't imagine why rockstars aren't beating a path to his door.

-----


It's still ludicrous. I'm not a civil engineer, but I was discussing this with my friends who are. In summary, the amount of damage that a vehicle does to a road is proportional to the fourth power of its weight.

A Nissan Leaf weighs 3,291 pounds (source: http://www.nissanusa.com/electric-cars/leaf/versions-specs/). A GMC Yukon Denali weighs 5,491 pounds (source: http://www.gmc.com/yukon-denali-luxury-suv/features-specs/ca...), or 1.67 times more. It should do approximately 7.75 more damage to the roads it drives over than the Leaf will.

At an average of 15,000 per year, the Yukon owner getting 16 mpg (source: average of 14mpg city and 18mpg highway from http://www.gmc.com/yukon-denali-luxury-suv/features-specs/po...) will burn about 950 gallons of gas per year. At $0.563 per gallon tax (source: combined state and federal taxes from http://en.wikipedia.org/wiki/Fuel_taxes_in_the_United_States), that owner will pay about $550 per year in taxes.

If NC is being honest about their reasons for taxing electric vehicle owners, then the Yukon owner is being subsidized about $225 a year (should pay: $775, will pay: $550) for driving a more polluting, more damaging vehicle. That's assuming that electricity is completely untaxed in NC so that the Leaf driver is paying only that $100 tax.

TL;DR This isn't about fundraising. It's about protecting gas-burning vehicle markets.

-----

bri3d 101 days ago | link

I do generally agree, but it's worth noting that in many states (including Colorado, a state which also charges a $50 EV fee), registration cost is based on vehicle weight and so the Yukon owner will pay a (still disproportionately low) greater amount.

-----

maxerickson 101 days ago | link

Roughly 0 people that are presently considering an electric car are concerned about $100 a year.

-----

sgustard 101 days ago | link

They might care because (1) a $100 tax can quickly grow to several times that, and (2) it's a slap in the face for doing the right thing.

-----

maxerickson 101 days ago | link

So do you think it is going to stop a lot of people from buying electric cars or are you just being picky?

Because I meant that it probably wasn't really going to stop very many people from buying electric cars (so it is then strange to assume it is about protecting the market for ICE cars).

-----

rbritton 101 days ago | link

In more northern states tires play a huge factor in road damage as well. The most noticeable road damage here in the Washington/Idaho/Montana area is from studded tires. The ruts from those tires are often an inch or more deep when they finally start repairs.

-----


White noise is an enormous source of entropy, by definition. For example, take a stream of samples and record whether the least significant bit is 0 or 1. That'd be almost impossible to predict or influence from outside the system.

-----


I've known people who made the same decision for many of the same reasons, and the universal experience I've heard was "it was all nice until we had to scale, and then it got expensive quickly". If your stack works for you, great! That's the important part! But if your traffic and income scaled by a factor of 100, would you be turning a profit or desperately looking for financing? You're the only one who can answer that, and it's something to consider.

-----

GFischer 182 days ago | link

Same as kstrauser: Microsoft products are great, but they're more expensive to scale.

I've been working with Microsoft products for close to a decade, and I think their software stack is great.

However, for my current side project, I went with another platform (Groovy+Grails, running on the JVM), purely because of cost (and also because I wanted to learn it :) ).

You can choose to be part of the BizSpark program and get all Microsoft software for free for 3 years, but you have to pay afterwards.

Prices for Microsoft software hosting is usually 2 or 3 times what it costs for a mostly equivalent software.

For example, something I'm using now (Amazon AWS), the price of a small Linux instance is $0,060/hr, while the equivalent Microsoft instance is $0,091/hr

Azure pricing is too expensive for me, I discarded it.

If I had money (maybe angel or VC backing), I might have considered it, but it's a bootstrapped side project.

-----

graycat 182 days ago | link

As far as I can tell from my software timings, likely revenue from ads as I need to scale, and Microsoft's prices, I should be able to afford Microsoft as I scale.

But at some point, if Microsoft's charges to me are so high I could hire a good team to refactor and convert my code to Linux and OSS, then I could save money. I have to believe that Microsoft has seen that issue before and has responses to it.

-----

More

Lists | RSS | Bookmarklet | Guidelines | FAQ | DMCA | News News | Feature Requests | Bugs | Y Combinator | Apply | Library

Search: