Topic: Closed - page 3. (Read 8662 times)

actually I have been digging quite deep into the bitcoin protocol, I can tell you now that I have spent a lot more than 100 hours making sense of how the protocol works. Even though Bouncy Castle did make life easy.

What makes a language "good" for bitcoin? .NET does a perfectly good job. I can make the same website in .NET as I can in PHP, in a less amount of time without having to pay for additional software.

(I really don't see the need for requiring bitcoin specific help for .NET when I understand the protocol myself and can implement the code myself. Which I would much rather do anyway, that way I can ensure the code is loosely coupled and I can write unit tests for it)
 
You are wrong. I have found it quite handy being able to generate bitcoin addresses, as well as using https://blockchain.info/pushtx to push transactions to the network. No need to use a bitcoind.

You are right in that it is NOT necessary, but it sure does help.  I would guess that you've never dug that deep into the Bitcoin protocol.  I have very much respect for anyone who can even understand it.  I've dedicated 100+ hours in trying to comprehend all of the crypto-jargon and its been a slow, uphill battle. My point being that if and when you need to customize / tweak your code (for some specific use case), unless you've got some sort of crypto-analyst (is that a real word?) on your team, you may find it hard pressed to achieve your goals.

BTW, the first link you sent is to a library that hasn't been updated since Oct '11 and the second was June '11. I'm sure at some point, someone thought it was a good idea to port that Java code, but it hasn't been touched since. And a .NET (dot net, ASP) search on SE came up quite skimpy (I count 2). I'm very sure there are ample and perfectly good uses for .NET, but Bitcoin is NOT one of them.

If you look at my other post I have pointed out a lot of security bugs in RoR and php has the least since it has the bigger amount of users. Also I would think python would be in the same boat as PHP and ASP.NET in with RoR.

Yes stackexchange is built on ASP.NET BUTTTTTTTTTTTTTTTTTT they have System engineers that have put a lot machines and hardware at the problem, maybe you should take a look at this

http://highscalability.com/blog/2009/8/5/stack-overflow-architecture.html

http://code.google.com/p/bitcoinsharp/ and http://bitcoincs.codeplex.com/

Also - I don't think it is necessary to have language specific support for bitcoin. There is a lot of information on how bitcoin works on the wiki as well as on http://bitcoin.stackexchange.com/

Since this thread derived a bit to security & implementation concepts, I don't think I'm too off topic and it reminded me of a use case for a smartcard wallet which might not be obvious to everybody - server side security.

You can support X * chips transactions per second on a server with a very good security level for cheap - an attacker will need to keep connected to the server in order to do something useful.



Of course if you plan to compete with Visa and Mastercard, old fashioned Hardware Security Modules are still recommended, but don't come with the same price tag 

(crossposted to my own thread with more details)

I would like to just chime in again.  As I this thread has made quite obvious, the framework, language, protocols that you enventually choose are very much debatable. When discussing "specifically" a Bitcoin business, which is what the OP has stated is his intention, I restate that security should be regarded as critical to your success. One bad hack and the results are disasterous as several Bitcoin business have already found out (the hard way). As much as I am enjoying the "friendly" banter, the point is sorely being missed here. Just go with what works for YOU.


I was unaware of Redmond's Bizspark program, as I got out of ASP well before 2008.  This is certainly a step in the right direction (a baby step at that), but nonetheless they are obviously trying to become more startup-friendly. I'd also like to mention that IE10 is again a great step in the right direction.  FINALLY, I mean FINALLY deciding to conform to the industry standards (so I don't have to build a website for EVERYONE and then IE).  If they make a version for Mac or Linux I think I'd actually be willing to install it and give it a test run.

Please correct me if I'm wrong, but I've found little to absolutely no ASP.net support when it comes to Bitcoin.  Now, I will say that I haven't really been looking so I could have easily overlooked that fact. Back to the point, IMO the OP will probably find a smoother road to success going with a Bitcoin-friendly platform/framework/language like PHP and/or Python (I don't think C/C++ is part of this debate) based on an open-source hardware infrastructure.  Things in the Bitcoin world ARE moving rapidly and I don't see it being easy relying on MSFT forums and support channels to get answers to your (Bitcoin security related) questions.

As ALL points are valid in their own right, considering the OP (and the Bitcoin-specific nature of this business), IMO the decision is quite clear:
(WARNING: troll-bait alert)

• NO to Windows v.Anything go with Linux v.AnyDistro
• NO to IIS - go with Apache or Nginx
• NO to ASP.net - go with PHP or Node.js
• NO to SQL Server - go with MySQL (not my first choice) or PostgreSQL or FirebirdSQL (my personal favorite), even MongoDB seems to be quite interesting

You won't have to worry about licensing fees EVER-EVER and support will be plentifully abundant in the FREE forums.

For this I offer 2 bitcents,
S.

No frameworks can have bugs that can hinder security so yes frameworks do dicate security. A language can be insecure making a website insecure, so your false, all three play a role in security.

Once again an entire post where you grabbed a bunch of random words and spewed it across the page without saying anything coherent.

1) You do realize that asp.net =/= SQL Server right?  You also know that Asp.net has database connectivity for every major (and lots of minor) RDMBS to include MySql?  Building asp.net application doesn't require the use of SQL Server.

2) On SQL Server vs MySQL scaling up isnt an issue.  All modern RDBMS can scale up.  Scaling out refers to using multiple servers to distribute the workload.  Get it OUT <-----> vs UP ^.     While this can be done with SQL Server the licensing costs generally making scaling UP more cost effective.   No I doubt you did realize that.  Nothing in the article talked about how "mysql would have worked but they were stuck with SQL Server".  MySql wouldn't have worked any better and unless it is an open source project MySql needs to be licensed.

3)
Yeah of course writing a custom RDBMS is something most startups are looking to do right?  Of course that custom RDBMS would also work with asp.net (and probably any other programming language).  DB =/= programming language.  Also how much cost (labor isn't free) do you think this custom RDBMS Twitter built ended up costing.  


4)
Lastly as we pointed out these are hardware scale issues way way way beyond what a startup would face.  ASP.NET and SQL Server do scale up into the "top of the web category" with sites like stackexchange which refutes your dubious claim that asp.net doesn't scale.   

Then again someone who finds a $300 conference "outrageous" likely hasn't had a very successful career in database development so don't beat yourself up for continually spewing nonsense.  I mean these are things picked up on the job and I doubt you will learn that stocking the shelves at Best Buy is very rewarding.

http://gigaom.com/2011/07/07/facebook-trapped-in-mysql-fate-worse-than-death/ - just saying.

No either you didn't read the link or you lack the knowledge to understand what you are reading.   Asp.net has data connectivity to a variety of RDBMS including MySql and Oracle.  However switching to MySQL wouldn't provide significnatly higher throughput on the same hardware and Oracle for the cost doesn't really make sense for the type of database they need.    The only thing which would give significantly better performance is a NO SQL setup like what Google uses but Stack Exchange didn't need that level of performance so the jump in complexity, and design using NO SQL wasn't warranted.   Maybe it would be someday if they scaled larger but given their "niche" scope it is unlikely they would ever need that level of performance so the huge code rewrite for NO SQL (not MySql) isn't warranted.   The one advantage that MySQL would have is that it easier to scaled out vs scaled up*.   Since it is more efficient when deploying SQL Server to scale up vs out that means making good hardware decisions. 

Of course we are talking about a scale of 20x to 100x larger than the largest Bitcoin enterprise.  The idea that this would be a problem for a startup is kinda laughable (it is a problem most startups wished the had). I would also point out that contrary to common knowledge MySQL is not license free unless the project is open source.   As many Bitcoin ventures are closed source they so require a MySQL license.

* Scale up would mean increasing the performance of a single (or small cluster) or database servers.  Where scale out would be replicating the database across a much larger cluster to achieve similar performance.  Since SQL Server is licensed the licensing costs are lower when scaling up vs scaling out.  The drop in server costs at the high end as well as moving storage to the SAN has made scale up less of a critical issue than in the past.   RAM has gotten a lot cheaper.  Building out a database server with quad xeons (32 cores) and 256GB or RAM as well as high end SAS controller (24x 2.5" backplane) is under $8K.   Going to 1TB of RAM, SSL offloading, and off server storage array is still under $10K.

Sorry....
I would worry about anyone with a masters degree in business... getting their business advice from people on an internet forum.
fine ask specific BITCOIN questions, but high security systems design is a completely different kettle of fish.

What point are you trying to get a across? That to optimise a website you need System Engineers? Well that's kind of a given when you are looking at a scale that big.

You will have these exact same problems with MySQL when looking at that scale. You can only scale up to some extent, then you will require techniques to allow you to scale out, like using replication.

That's why I added, "for the most part". Please point out any security issues you are aware of with PHP, ASP.NET, Python or Ruby on Rails.

Yeah nothing in there was right.  Not scalable? Stackexchange is an example of a web project built using ASP.NET (2.8 million users, 13 million+ questions & answers)
As for licensing startups can get licenses for up to three years using Microsoft bizspark program at no cost.
Microsoft pulling the plug on asp.net or MVC?  Absolute nonsense.  

Just the normal mindless "Microsoft sucks" from people who have never done any professional software development in their lives.
 

For the most part, a framework doesn't dictate the security. A language/platform doesn't usually make a website "insecure", the code or configuration usually does.

You say Rails is awesome but no one has given hard facts as to why?

THAT is the truth right there .. take note.


Um, this is going to be a Bitcoin business. Personally, I think security should #1 AND #2. In the end, who gives a damn how many times your server crashes in a day, if ALL your Bitcoins magically disappear. I would strongly advise that you separate your web server from your bitcoin server as was suggested earlier. And at the very least put some IDS software on so you know what the hell is happening at all times:

• OSSEC is FREE and well supported http://www.ossec.net/
• Snorby is pretty f---in' awesome, but its RoR and I (don't know why, but) just don't like RoR https://snorby.org/

I have extensive experience working with both ASP.NET and PHP and if startup/operational costs matter at all to you, I'd say choose anything BUT ASP.NET. Relatively speaking (and I acknowledge there is obviously some value from Redmond) they are just WAY, WAY too expensive (licensing every which way you turn -- till you get dizzy and collapse). Unfortunately, many young startups find this out too late.

Just my 2 bitcents
S.

Not offering advice - just facts.   

Who is going to take the advice of a man who's icons is a fat nibble licker.

Thanks. but no thanks.