Timeline for avoiding multiple calls to SQL while being modular
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 12, 2014 at 23:28 | history | edited | Arseni Mourzenko | CC BY-SA 3.0 |
edited body
|
| Oct 12, 2014 at 21:55 | comment | added | Arseni Mourzenko | @user20358: I also often just feel that something will slow my code. Then I do profiling and I discover that the bottleneck is completely unrelated. Or I do the change and discover that the code is now 0.001% faster (or maybe 60% slower). That's why premature optimization is evil, since it makes you write code which is not as clean as it should be, with no actual benefit. | |
| Oct 12, 2014 at 19:52 | comment | added | user20358 | I've really not been able to profile the application because we are still in the design phase. I just feel that making three trips to a server when one trip can do the job is taking too many liberties too early. I could be wrong on this thought and Im willing to listen to critcism on this point. | |
| Oct 12, 2014 at 19:50 | comment | added | Arseni Mourzenko | @user20358: then you're at a point of doing premature optimization. As for your question about non-functional requirements, see that answer, point 2 for an example of such requirement. If instead you'are wandering what is the usefulness of writing down requirements before developing a software product, than this may be a separate question to post on Programmers.SE. | |
| Oct 12, 2014 at 19:45 | comment | added | Arseni Mourzenko |
@randomA: with TransactionScope in .NET Framework 4, it doesn't matter, since you can mix for example several queries to the database with some calls to NTFS with transactional support, all this within a single transaction.
|
|
| Oct 12, 2014 at 19:42 | comment | added | user20358 | the application which I described above is not yet written. At this point in time the possibility of individual views being required is not yet 100% But they are being written just because eventually they will be required as per the requirements we know today, which could also change in the future. what exactly did you mean by writing down a non-functional requirement ? | |
| Oct 12, 2014 at 17:33 | comment | added | InformedA | Very well written answer, I just want to add that you also might want to mention the correctness aspect of the question. If one calls multiple non-DB operations, I am not sure how one can use the DB's built in transaction support. | |
| Oct 12, 2014 at 17:05 | history | edited | Arseni Mourzenko | CC BY-SA 3.0 |
added 1300 characters in body
|
| Oct 12, 2014 at 16:50 | history | edited | Arseni Mourzenko | CC BY-SA 3.0 |
added 1300 characters in body
|
| Oct 12, 2014 at 16:41 | history | answered | Arseni Mourzenko | CC BY-SA 3.0 |