J. Clark. Expat - XML parser toolkit. http://www.jclark.com/xml/ expat.html, 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Since n is bounded by the height of the XML data tree and m by the small amount memory allocated to store the Markov histogram, m and n are practically constants. Therefore the update overhead is a constant. 5 Experiments We implemented our XPathLearner in C using the XML Parser Toolkit [7]. We describe in this section the experiments that we have done to validate our method. We describe briefly the data sets used, the query workloads, the performance measures, and the goals of the different sets of experiments. Performance Measures. Two error metrics were used: the average ....

J. Clark. expat--XML parser toolkit, 2000.

.... false; member(X, X Xs] true; member(X, Y Ys] member(X,Ys) Functions are grouped into modules, and a subset of those functions can be exported, declaring both function name and arity, to be used by other modules. module(mymod) export( member 2] Hence, mymod:member(3, [1,2,3]) reduces (i.e. evaluates) to the atom true, while mymod: member(5, 1,2,3] reduces to false. Besides lists, Erlang programmers also can use tuples, similar to C structures, which are constructed in arbitrary but finite length by writing fA,B, Cg. Lists and tuples can hold any valid Erlang ....

....Functions are grouped into modules, and a subset of those functions can be exported, declaring both function name and arity, to be used by other modules. module(mymod) export( member 2] Hence, mymod:member(3, 1,2,3] reduces (i.e. evaluates) to the atom true, while mymod: member(5, [1,2,3]) reduces to false. Besides lists, Erlang programmers also can use tuples, similar to C structures, which are constructed in arbitrary but finite length by writing fA,B, Cg. Lists and tuples can hold any valid Erlang value, from numbers and atoms to lists, tuples and even functional values. As ....

[Article contains additional citation context not shown here]

J. Clark. expat - XML Parser Toolkit 1.1. http://www. jclark.com/xml/expat.html, 1999.

....gives signi cantly larger network transmission times for XML messages, increasing latency and decreasing possible message rates substantially. 3 A variety of implementations of XML, including both XML generators and parsers, are available. We have used the fastest known to us at this time, Expat [25]. 13 The NDR approach signi cantly improves sending side costs. As is mentioned in Section 3, we transmit data in the native format of the sender. As a result, no copies or data conversions are necessary to prepare simple structure data for transmission. So, while MPICH s costs to prepare for ....

....signi cant exibility in reacting to changes in the incoming data and was our initial choice when implementing NDR. XML necessarily takes a di erent approach to receiver side decoding. Because the wire format is a continuous string, XML is parsed at the receiving end. The Expat XML parser [25] calls handler routines for every data element in the XML stream. That handler can interpret the element name, convert the data value from a string to the appropriate binary type and store it in the appropriate place. This exibility makes XML extremely robust to changes in the incoming record. ....

J. Clark, \expat - XML parser toolkit." http://www.jclark.com/xml/expat.html.

....in High Performance Systems 7 milliseconds (logscale) 100Kb 10Kb 1Kb 100b 100.0 10.0 1.0 Data Structure Size .01 1000.0 .1 MPICH CORBA ECho XML Receiver Decode Times Figure 5. Receiver side costs for XML, MPI and NDR. Logarithmic scale used in vertical dimension) XML parser 3 [5] calls handler routines for every data element in the XML stream. That handler can interpret the element name, convert the data value from a string to the appropriate binary type and store it in the appropriate place. This flexibility makes XML extremely robust to changes in the incoming record. ....

....rather than reusing the receive buffer (as we do) However, ECho s receiver side conversion costs still contribute roughly 20 of the cost 3 A variety of implementations of XML, including both XML generators and parsers, are available. We have used the fastest known to us at this time, Expat [5]. milliseconds (logscale) 100b 1Kb 10Kb 100Kb 10.0 1.0 .1 Data Structure Size .01 ECho DCG ECho MPICH Receiver Decode Times Figure 6. Receiver side costs for interpreted conversions in MPI and ECho and DCG conversions in ECho. of an end to end message exchange. While a portion of ....

James Clark. expat - XML parser toolkit. http://www.jclark.com/xml/expat.html.

....3 3.1 Overview Figure 4 presents the actors involved in the Erlatron subsystem. Erlatron adapts the Sablotron library by using a port which performs basic transformations of a couple of Erlang binaries representing the XSL stylesheet and the XML source. Sablotron is based on the Expat library [3] and has been designed for performing fast and compact transformations. The port is managed by a generic server that o ers XSLT services to any client. This slave server constitutes the basic processing unit and, considering the CPU cost of performing XSL transformations, low additional overhead ....

J. Clark. Expat - XML Parser Toolkit 1.1. http://www.jclark.com/xml/expat.html.

....al. 1998) it seems evident to base the metadata representation on the generic XML standard, rather than the specific format originally created for Busstop. This would also reduce the size of the code base dramatically, as all parsing tasks could be moved to a generic XML library, such as expat (Clark, 1998). Bush is another part to reconsider seriously: while useful, it really is somewhat limited in terms of expressivity, compared with the more and more widespread scripting languages, such as Python (Rossum, 1995, Rossum, 1999) and Perl (Wall et al. 1996) One possibility could be to repackage ....

Clark, J., 1998. expat---XML parser toolkit. http://www.jclark.com/xml/expat.html.

....2 , 1 The Sun machine is an Ultra 30 with a 247 MHz cpu running Solaris 7. The x86 machine is a 450 MHz Pentium II, also running Solaris 7. 2 A variety of implementations of XML, including both XML generators and parsers, are available. We have used the fastest known to us at this time, Expat [3]. 3 sparc encode network i86 decode i86 encode network sparc decode .034m .227m .063m .010m .227m .104m sparc encode network i86 decode i86 encode network sparc decode .086m .345m .106m .046m .345m .186m sparc encode network i86 decode i86 encode network sparc decode .971m 1.94m 1.19m .876m ....

....significant flexibility in reacting to changes in the incoming data and was our initial choice when implementing NDR. XML necessarily takes a different approach to receiverside decoding. Because the wire format is a continuous string, XML is parsed at the receiving end. The Expat XML parser[3] calls handler routines for every data element in the XML stream. That handler can interpret the element name, convert the data value from a string to the appropriate binary type and store it in the appropriate place. This flexibility makes XML extremely robust to changes in the incoming record. ....

J. Clark. expat - xml parser toolkit. http://www.jclark.com/ xml/expat.html.

....Filter Engine, it is run through an XML Parser which then drives the process of checking for matching profiles in the Index. We use an XML parser that is based on the SAX interface, which is a standard interface for event based XML parsing [Meg98] We developed the parser using the expat toolkit [Cla99a], which is a non validating XML processor. The SAX event based interface reports parsing events (such as encountering the start or end tag of an element) directly to the application through callbacks, and does not usually build an internal tree. To use the SAX interface, the application must ....

J. Clark, "expat - XML Parser Toolkit", http://www.jclark.com/xml/expat.html, 1999.

....measured by replacing the network send( routine with a null operation. The XML costs represent the processing necessary to convert the data 2 A variety of implementations of XML, including both XML generators and parsers, are available. We have used the fastest known to us at this time, Expat [3]. 1Kb 10Kb 100Kb 1000.0 .001 100.0 10.0 1.0 100b .01 .1 XML PBIO MPICH Message size Sender Encode Times Figure 3: Send side data encoding times. from binary to string form and to copy the element begin end blocks into the output string. Just one end of the encoding time for XML is ....

....receiver side decoding. Because the wire format is a 100b 1Kb 10Kb 100Kb 100.0 10.0 1000.0 binary data size .01 .1 1.0 Receiver Decode Times XML MPICH PBIO Figure 4: Receiver side costs for XML, MPI and PBIO. continuous string, XML is parsed at the receiving end. The Expat XML parser[3] calls handler routines for every data element in the XML stream. That handler can interpret the element name, convert the data value from a string to the appropriate binary type and store it in the appropriate place. This flexibility makes XML extremely robust to changes in the incoming record. ....

James Clark. expat - xml parser toolkit. http://www.jclark.com/xml/expat.html.

....1 The Sun machine is an Ultra 30 with a 247 MHz cpu running Solaris 7. The x86 machine is a 450 MHz Pentium II, also running Solaris 7. 2 A variety of implementations of XML, including both XML generators and parsers, are available. We have used the fastest known to us at this time, Expat [3]. 5 1Kb 100Kb 1000.0 .001 100.0 10.0 1.0 .1 .01 10Kb 100b CORBA PBIO MPICH Sender Encode Times binary data size XML (a) Send side encoding times binary data size 100Kb 10Kb 1Kb 100b 100.0 10.0 1.0 .1 .01 1000.0 PBIO Receiver Decode Times CORBA MPICH XML (b) Receive ....

....significant flexibility in reacting to changes in the incoming data and was our initial choice when implementing NDR. XML necessarily takes a different approach to receiver side decoding. Because the wire format is a continuous string, XML is parsed at the receiving end. The Expat XML parser[3] calls handler routines for every data element in the XML stream. That handler can interpret the element name, 7 binary data size 1Kb 10Kb 100Kb milliseconds (logscale) 10.0 100b .01 .1 1.0 PBIO DCG PBIO MPICH Receiver Decode Times Figure 3: Receiver side costs for interpreted ....

James Clark. expat - xml parser toolkit. http://www.jclark.com/xml/expat.html.

No context found.

J. Clark. Expat - XML parser toolkit. http://www.jclark.com/xml/ expat.html, 1998.

No context found.

J. Clark. Expat - XML parser toolkit. http://www.jclark.com/xml/expat.html, 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC