meta user
network profile
| bio | website | |
|---|---|---|
| location | Los Angeles, CA | |
| age | ||
| visits | member for | 2 years, 8 months |
| seen | 10 hours ago | |
| stats | profile views | 194 |
Before The Subversion, there was either The Control, or The Chaos. There was no Complaint. Everything was as peaceful as it could be.
|
May 12 |
comment |
When to Use workflow engines? @JamesAnderson: except the customer can now pay their own non-programmers (workflow consultants) to troubleshoot their own rules, without having to file a support case with the software vendor. |
|
May 12 |
comment |
Keyword Collection, ManyToOne or OneToMany relationship? See ManyToMany bidirectional eager fetching. stackoverflow.com/questions/3978733/… |
|
May 11 |
comment |
How to simplify my complex stateful classes and their testing? Related: sharing large number of states and available actions between A.I. agents in computer game F.E.A.R. docs.google.com/viewer?url=http://web.media.mit.edu/~jorkin/… |
|
May 11 |
comment |
Contributing to an open source project It takes time. It might take years of time. |
|
May 11 |
comment |
Language Learning: Well documented C projects? Have you taken operating systems, digital logic design (which covers state machines), and embedded systems in school? These courses might be essential prerequisites. |
|
May 10 |
revised |
Is RESTful API appropriate for services which are nearly computational in nature? added RabbitMQ idea page for REST API |
|
May 10 |
comment |
How to model/structure data that contains parent/child relationships for the same entity? Can a question have multiple parents? Is there a depth limit to parent-child relationship? |
|
May 9 |
comment |
What does Uncle Bob mean by 'noun phrase names'? maybe Uncle Bob is adopting RESTful thinking in extolling the virtues of resource-as-objects. |
|
May 9 |
comment |
moving from development to quality assurance Might also consider IT / SysAdmin. |
|
May 9 |
revised |
Statistics collection engine for C++ systems added 131 characters in body |
|
May 9 |
comment |
Data structure for accessing units of measure Similar issue occurs in DateTime and TimeSpan in C#. One suggestion is to include an Origin field in the class. To encode zero Celsius (freezing point) the Value field would be set to 0 Kelvin, and the Origin field would be set to +273.16 Kelvin. Converting to Kelvin requires 3 steps: un-shift, scale, re-shift. To encode a temperature difference of 5 degrees Celcius, the Value field would be set to 5 K and the Origin field would be set to 0 K, meaning that shifting should not be performed when converting to Kelvin. |
|
May 9 |
answered | Data structure for accessing units of measure |
|
May 9 |
comment |
Data structure for accessing units of measure Is the application sensitive to rounding errors? Can the quantities be converted back and forth to the SI system as many times as needed? Also, how many distinct physical quantities are needed (including derived and rate units)? |
|
May 9 |
revised |
Is RESTful API appropriate for services which are nearly computational in nature? edited tags |
|
May 9 |
asked | Is RESTful API appropriate for services which are nearly computational in nature? |
|
May 9 |
comment |
Heap overflow vs stack overflow Keep in mind there are newer VMs in which the program stack space is allowed to be fragmented. In other words, each function's call frame is dynamically allocated from heap. For these VMs, the distinction between heap space and stack space disappear. |
|
May 9 |
comment |
Heap overflow vs stack overflow heap space can be fragmented but stack space (per thread) is always contiguous and has a fixed starting point. The corollary is then when the growth of the stack space hits the first unmovable object in memory, the growth fails. In addition, on some OSes e.g. Windows each thread is typically assigned a small stack limit (e.g. 0.5 MiB) which cannot be grown. |
|
May 9 |
comment |
Heap overflow vs stack overflow @Rayhunter: For 32-bit systems, this situation can (and should) be avoided by using memory-mapped file, and by modifying the algorithms to perform the work incrementally (touching a small section of the data at a time - small enough to fit in memory). For 64-bit systems, exhaustion of virtual address space is nearly unheard of. |
|
May 9 |
comment |
Which design pattern would be best for this case? You can try to decompose the code into a number of "code blocks", and then see if next year's policy can be composed from these blocks. However, if next year's policy turns out to be unforeseeable (requiring completely new code blocks and behavior), code change is inevitable. |
|
May 9 |
revised |
Statistics collection engine for C++ systems added 457 characters in body |
