Thread 3

The Free Lunch Is Over

Optimization the word that engineers from different disciplines like to hear. There is nothing more accelerating and frustrating at the same time, than having a working solution for a problem and someone come to you and say: "I would like this to be better". Is a challenge, trying to abstract the problem and thing of ways of making it "better". In programming this could be so many things, make the program faster, requires less space in memory, requires less petitions of IO. So many ways that a program can become better, yet as soon as I mentioned some of the solutions for a program to be better problems came out. A program cannot improve in every aspect without some fallback. Trying to make an algorithm faster can cause a larger utilization on resources. Less petitions to IO means that we require to save the information somewhere so we can consult it later. This is where knowing the type of problem we are presented and analyzing the proper solution is an important subject. Herb Sutter gives good points and that any programmer should keep in mind when programming a new solution. Just that i think something is being happening this couple of years that has revolutionize the way we provide solutions to our clients, Cloud computing. Herb Sutter is worried about how we are not able to rely on computer components in becoming faster and faster so our algorithms can run faster. This is completely true, yet the company that offers the fastest solution on the market , IBM, is not the one dominating on solutions for processing petitions to different servers. Amazon found an amazing product that is dominating the market that any person can use without any problem. People are becoming experts on the platform and understanding different services that can solve the issue that Herb Sutter mentioned. Is easy creating new instances of a service on amazon and implementing a load balancer that could take the different operations and delegated them to the different instances. Now this doesn't give us the lunch back, in reality we should know which kind of hardware solution we will be using and where our software would be deployed and exploit the benefit of each architecture. If this is a distributed solution as Amazon or a Mainframe with IBM. 

Comments

Post a Comment

Popular posts from this blog

Thread 1

This is the first Post for the multithreading programming class, Hi my name is Arturo Velazquez and I am interested in all the things that are related to some kind of development. Right now I work for an American company called LendingClub as a front end developer. What I expect from this subject is to improve my skils in programming for C++. C++ is a language that i want to dominate and be fluent in with. I have different hobbies when I have time , like playing video games and miniatures.
Read more

Thread 4

The trends for programming will always keep changing doesn't matter in which sector we are focused on. We have seen it through out the years on how different ways of programming have become more relevant in different periods of time. Now the reason for why a trend becomes popular could be for a number of reasons. This could be for development of better technology, easier understanding on the programming part of the development or even the demand of developers on certain language. In this case knowing that parallelism is going to become the next trend that programmers will need to learn. Almost all computers and processors have a dual core integrated to them. So is easy to see that people will try to improve their execution times of programs by using multiprocessors. I think this is becoming more relevant with a new way of computing. Quantum computing will cause a whole revolution on how we understand and solve computer problems. A major problem that will cause with this revolution
Read more

Thread 8

The Erlang programming language is an interesting approach to a solution for concurrent and parallel programming. It is interesting for so many different aspects and the first aspect would be that is a programming language that doesn't require the idea of sharing memory. This is a big change in paradigms , we have seen it in some old languages that the idea to share memory is a common one. Yet even if it is an old idea and a paradigm that is implemented in several languages , is hard to handle. It is true that we have certain objects that could help us to manage the memory and how different threads or programs can access shared resources. This sometimes can be difficult to understand and even sometimes hide some bugs that are hard to replicate and even solve them.  Another idea that Erlang deviates from is the idea of immutable variables or states. This is another old way to manage resources and make sure that we can keep the state of the program appropriately. We can notice that E
Read more