recentpopularlog in

mcherm : via:debasishghosh   17

It's the familiarity model! - Ruminations of a Programmer
Sucinct code and incomprehensible code are the same thing; the difference is whether the programmer is familiar with the idiom.
programming  languagedesign  DebasishGhosh  via:DebasishGhosh 
february 2012 by mcherm
Ruminations of a Programmer: Using generalized type constraints - How to remove code with Scala 2.8
How scala allows you to add a method to a parameterized class (eg: collection of Foo) that can only be called (checked at compile time not run time) if the parameter (Foo) meets certain criteria. For instance, add a "sum" method on Collection that only works if the contents are Numbers.
scala  languagedesign  via:DebasishGhosh  types 
august 2010 by mcherm
Worst Case Complexity « Gödel’s Lost Letter and P=NP
A possible alternate means for analyzing algorithms: based, not on WORST case (which is unreasonably bad) or average case (which may be too good) but on an opponent with only polynomial-time resources available.
complexity  computerscience  math  algorithms  via:DebasishGhosh 
june 2010 by mcherm
Ruminations of a Programmer: Why I don't like ActiveRecord for Domain Model Persistence
Having domain objects automatically persisted is nice. But having to extend some class to do it (ActiveRecord) is not so great. For one thing, it ties you down to just one persistence mechanism. Debasish Ghosh has a solution that's less intrusive (basically, it uses introspection).
database  DebasishGhosh  via:DebasishGhosh  programming  scala  activerecord  persistence 
february 2010 by mcherm
actorom - Google Code
A library for actors in Java. Among other things, it struck me as having good documentation.
java  library  concurrency  programming  via:DebasishGhosh 
july 2009 by mcherm
How laziness changes thinking in Haskell - Lost in Technopolis
Some interesting examples of how it's difficult to debug in Haskell because of the laziness.
haskell  functional  laziness  programming  via:DebasishGhosh 
april 2009 by mcherm
Your data model can speak different languages too: Ruminations of a Programmer
Some thoughts on non-traditional forms of data storage (anything other than a RDBMS), and why you might want them, especially if extreme scalability is an issue.
via:DebasishGhosh  DebasishGhosh  programming  scalability  database  modeling 
march 2009 by mcherm
Bloom Filters - optimizing network bandwidth: Ruminations of a Programmer
Another great idea: transfer a Bloom filter over the network and use it to pre-filter the data that's going to be merged together (perhaps when doing a map-reduce for instance). Then you can send a much smaller amount of data over the network to do the actual merge.
bloomfilter  programming  mapreduce  datastructures  via:DebasishGhosh  DebasishGhosh  concurrency 
march 2009 by mcherm
My Biased Coin: Hash Table with Preemptive Bloom Filter
Ah... if your hash table lookups take much time, use a bloom filter in front to do them quickly. (But why is the lookup so slow? Is the table too big to fit in memory?)
datastructures  algorithms  via:DebasishGhosh 
march 2009 by mcherm
too much TemplateMethod « silk and spinach
I wish I had this much insight into my own programming peccadilloes.
via:DebasishGhosh  programming  patterns 
january 2009 by mcherm
A Rant A Day - The Power of The Functional or: I’ll bet you can’t do this in Java
All simple examples, but this illustrates some basic things you can do in a language with first class functions.
functional  programming  via:DebasishGhosh 
january 2009 by mcherm
Ruminations of a Programmer: Data 2.0 - Is your Application ready ?
The idea of a big database that runs on a single large machine ought to die. Start by assuming that everything has to scale larger than one machine -- and only use databases that are designed to work this way.
scalability  database  via:DebasishGhosh  programming 
october 2008 by mcherm
Towards Robust Distributed Systems: PODC-keynote.pdf (application/pdf Object)
Inktomi runs a data center that needs 100% uptime... through software and OS upgrades, even through data center moves. Here are some lessons from that, which include considering the impact of the CAP theorem (consistency, availability, and resistance to partitions: can't have all three).
scalability  highavailability  programming  computerscience  via:DebasishGhosh 
august 2008 by mcherm
Scaling Out | Engineering @ Facebook's Notes
When Facebook wanted a second data center with its own DB, how did they need to change their application? (1) make one DB read-only and replacate from the main datacenter; modify MySQL to clear memcache on replication. (isn't open source cool?) (2) all pages doing edits go to main data center. For the first few seconds after a user makes a change, send that user to main data center as the other may still have the old value.
scalability  JasonSobel  memcache  opensource  mysql  mypostings  concurrency  parallelprogramming  via:DebasishGhosh  programming 
august 2008 by mcherm

Copy this bookmark:

to read