Is TimesTen from Oracle adequate? It really is about 10 times faster than stock Oracle.

<t>QuoteOriginally posted by: MarineQuoteDDE Advise is asynchronous, where RTD is synchronousRTD is asynchronous as well, I am not sure where you got the idea that it wasn't. RTD appears asynchronous to the user, but the underlying communication mechanism is synchronous. All COM calls are synchronou...

<t>You could use Cascade DataHub to move data between Excel and web browser, between Excel and Excel on a network or between custom applications and Excel. It uses DDE and TCP as the underlying mechanisms. I know that people here generally claim that RTD is better, but it isn't clear-cut. For single...

<t>The password file is world-readable by definition. User programs without special permissions sometimes need to access this information. Shadow is not world-readable. This means that you need extra permissions to be able to read it. This stops joe average from performing a brute-force cracking job...

<t>The only open-source data server/client architectures I'm aware of are for creating games. That's probably not what you want, but you can search sourceforge if you that interests you. In the commercial world there is Cascade DataHub. It has some of the properties you describe. It's used by trader...

<t>You could use Cascade DataHub running in Linux (click here), use a shell script or Perl script running in a timed loop to read data from /proc and populate whatever data you want into Cascade DataHub. Connect using an SSH tunnel securely from anywhere on the Internet using Cascade DataHub running...

<t>QuoteOriginally posted by: farmerQuoteOriginally posted by: gcYou can have a single-threaded program that waits on many resources and gets awaken by the kernel when each of those resources become available again. UNIX C call is select().I use select in my single main thread, but doesn't it block?...

<t>QuoteOriginally posted by: farmerI can guarantee plenty of sections of code snippets take less than a certain amount of time. And so I don't need to slice them up any further. It's not like my neighbor is running all kinds of crap on my computer I don't know about. If a computer is programmed to ...

<t>I would use SVN over CVS, by a small margin. I like the Windows based integration of SVN, and as projects get larger SVN takes a clear lead in performance. SVN also does a better job of handling binary files, in my opinion. Annoyance of SVN: it enforces line termination consistency, so if you wor...

<r>You can collect tick data via DDE from Bloomberg or Reuters using Cascade DataHub (<URL url="http://www.cogent.ca">www.cogent.ca</URL>) and then write a simple script that saves the data to file. Or, you can put the data into Excel from your stock feed and then collect it from Excel using Cascade...

<r>That depends on your feed. You could try using Cascade DataHub at <URL url="http://www.cogent.ca">www.cogent.ca</URL>. It converts among DDE, TCP and ODBC and includes a TCP API for C++, .Net and Java. If your feed is DDE, or from Excel, then it will work. If your feed offers a custom API in C++ ...

<t>Quotefor (int j=0;j<savedArray.Length;++j){savedArray[j] = (byte) (savedArray[j] >> 4);savedArray[j] = (byte) (savedArray[j] << 4);}Do you really need to compute savedArray.Length every time? Also, removing the last 4 bits of an integer is an & operation. It's one of the fastest integer opera...

<t>One possible solution, depending on the type of results you get from Matlab:Cascade DataHub is a data middleware product that can make Matlab calls and put them into a name/value data set. That data set can then be used to populate a web page either through AJAX or Java. You can also use the Casc...

<t>QuoteOriginally posted by: weismatDepending what you do, Vista can be interesting, because MS has rewritten the TCP implementation. This is very big for realtime-data processing.What do you mean by that? I thought they just added IPV6. MS's TCP/IP implementation has always been able to move data ...

<t>It sounds to me like you need your XLL to contain an interpreter. Consider embedding a language like LUA into your XLL and then have your web service submit LUA code to it. Then create 100 (or 1000, whatever) generically named Excel-callable wrapper functions, each of which simply looks up an ass...