id
stringlengths
7
12
sentence1
stringlengths
5
1.44k
sentence2
stringlengths
6
2.06k
label
stringclasses
4 values
domain
stringclasses
5 values
train_126600
Since both prices and quantity are the same as in the two-sided Cournot game, payoffs are also unchanged.
f is an isomorphism.
reasoning
Software and its engineering
train_126601
The Least Squares technique is not robust in the face of outliers, such as those in heavy tailed distributions, while Least Absolute Deviations is.
the Least Absolute Deviations is more applicable for use with software metrics data.
reasoning
Software and its engineering
train_126602
This is due to the fact that this systematic mapping focus is on investigation of tools and models that utilize ontologies as a support to distributed software development activities.
it is important to take under consideration that most of the resources allocated in this branch have also been included in other knowledge branches.
reasoning
Software and its engineering
train_126603
Ad­ditionally, the class that becomes a subclass in an inheri­tance relationship can only be tested in the presence of itssuperclass or a driver.
in an example as the onepresented in Figure 7 the ORD generated consists of twointertype relationships and an inheritance relationship.
reasoning
Software and its engineering
train_126604
Since the func­tions in a declarative program are typically very small and consist of a few lines only, it is not suf.cient in practice to cover only the branches of the function to be tested.
we will aim at covering the branches of all the directly and indirectly called functions, too.
reasoning
Software and its engineering
train_126605
Despite the above observations, the early activities in AD, like deciding on the coarse-grained architecture, are currently mostly guided by the intuition and the experience of the software architects.
those activities tend to be unsystematic, hard to predict or plan and difficult to perform by novices in architectural design.
reasoning
Software and its engineering
train_126606
*/) { state_ = 4; 29 30 } else { state_ = 5; 31 32 } break; 33 case 4: 34 35 /* send message m3 */ 36 // recv message m4 37 ACLMessage msg = myAgent.receive(); 38 Object content = msg.getContent(); 39 40 /* handle the content */ 41 state_ = 3; 42 break; 43 case 5: 44 /* send message m5 */ 45 break; 46 case 6: 47 state_ = 0; 48 break; 49 case 7: 50 /* send message m7 */ 51 isTerminate_ = true; 52 break; 53 54 55 } } public boolean done() { 56 return isTerminate_; 57 585 } } Figure 2: An implementation of Role1 in JADE to modify both of them if some changes to the message exchanges are to be applied.
maintainability of interactions decreases.
reasoning
Software and its engineering
train_126607
On the other hand, removing the element from the subsystem again does not decrease its streets length.
even currently small subsystems might be represented by long streets which have negative impact on the layout compactness.
reasoning
Software and its engineering
train_126608
Participants 1Twitter (http://twitter.com) has been designed as a pub­ lic communication service.
users do not assume their conversations private.
reasoning
Software and its engineering
train_126609
In addition, a series of small changes from version 3 and 6 might contain fixes to the defects made in version 3, so comparing 3 and 6 can help identifying defects.
4 and 5 might be excluded from the initial inspection.
reasoning
Software and its engineering
train_126610
Such an approach might be an example were the diversity in the versions to be true diversity.28 What we are concerned about is keeping the autopoietic part of the overall system working.
it is not correctness of result that is important but survivability.
reasoning
Software and its engineering
train_126611
With static scheduling, where every thread must perform the same amount of work, performance varies from 11.4x speedup with the high tax rate to 8.8x speedup with a low tax rate, because of load imbalance while the N-Queen is running.
the use of tax must consider both the bene.t to sequential programs and the cost to the parallel programs that are preempted.
reasoning
Software and its engineering
train_126612
Because of the network delay between the sending and receiving players, when a DR vector is received and rendered at a player, the original trajectory of the player/entity may have al­ready changed.
in physical time, there is a deviation at the receiving player between the exported path and the rendered tra­jectory (which we refer to as placed path).
reasoning
Software and its engineering
train_126613
We normalized the drift and volatility to set T = 1.
a time frame of 5 years or 20 quarters yielded a drift rate of µT = -0.572, and a voliatility svT= 0.746.
reasoning
Software and its engineering
train_126614
4.1 Basic WUIs In order to support the construction of WUIs for a large class of applications, we only require that the application program supplies the WUI with the current state of the data and an operation to store the data modi.ed by the user.
the main operation to construct a WUI could have the type signature (remember that dynamic web pages are of type IO HtmlForm) mainWUI :: a -> (a -> IO HtmlForm) -> IO HtmlForm so that an expression of the form (mainWUI dstore) evaluates to a web page containing an editor that shows the current data d and executes (store d') when the user submits the modi.ed data d'.
reasoning
Software and its engineering
train_126615
Taking into account these assumptions, the probability that all n assignments in the sample are contained in the lower (100 - P )% of the population is computed as: P (AI)=( 100-P )n. As A and 100 AI are opposite events, the sum of probabilities that they occur is equal to 1: P (A)+ P (AI)=1.
the probability of the event A can be computed as: I 100-P n P (A)=1 - P (A)=1 - 100 We observe that the probability that a sample of random task assignments contains at least one of the P % of the best-performing assignment is independent of the population size (i.e.
reasoning
Software and its engineering
train_126616
Unfortunately, the community has not yet developed a mul­tiprocessor scheduling algorithm with .nite speed compet­itive ratio for tasks that share resources on uniform mul­tiprocessors.
in this paper, we present one and prove its performance.
reasoning
Software and its engineering
train_126617
The .elds() method of the View feature uses its parameter as the types of the .elds that will be returned.
it is possible to use the method in both ArrayList and LinkedList generators by passing the appropriate type as the argument.
reasoning
Software and its engineering
train_126618
When d =3, the 3-0, 1-2, and 0-3 cases all entail recursive calls, but only the 2-1 case causes a defective item to be deferred.
3p2q 2 D3 = p3D3 +3pq +q3D3 = = p. 1 - p3 - q3 For d = 4, the 4-0 and 0-4 cases both entail recursive calls, the 3-1 case entails a 3-defective recursive call and 1 deferral, the 2­2 case entails 2 deferrals, and the 1-3 case entails a 3-defective recursive call.
reasoning
Software and its engineering
train_126619
At a particular program statement s, the predicate transparent(E)should be true iff the value of the pro­gram expression in the meta-variable E cannot be affected by s.We take the obvious (but slightly conservative) approach of de.ning this to mean that none of the program variables occurring in the expression are changed by s, so for example a :=3 is transparent to b +c, but not to a +c (ignoring aliasing effects).
for the predicate transparent(E)and the statement a :=3, we need to gen­erate a constraint for E that allows it to take on all possible program expressions that do not contain a.
reasoning
Software and its engineering
train_126620
Since LDPC code can e.ectively correct errors, the length of wa­termarks or codewords can be minimized.
im­ages can retain high levels of quality.
reasoning
Software and its engineering
train_126621
Each execution of repOk re-initializes the state of the structure.
the structure is constructed with every repair action, which adds a quadratic effect on the run­time of STARC.
reasoning
Software and its engineering
train_126622
Therefore, the open-ended structure of the system was of paramount importance.
the front end was to provide information about the program that would be sufficient to support a wide class of SAT methods.
reasoning
Software and its engineering
train_126623
Their travel patterns are highly predicable and regular.
there is an opportunity to form recurring virtual mobile communication networks and social communities between these travelers or their vehicles in the form of vehicular social networks (VSNs) [1].
reasoning
Software and its engineering
train_126624
2.2 Nonuniform Quantization Schemes The .nite word-length of the soft information directly de­cides the size of memory, which plays a role to store the intrinsic and extrinsic messages and dominantly determines the overall implementation area in the partially parallel LDPC decoder.
the reduction of .nite word-length with­out signi.cant performance loss can decrease the hardware size which includes the computation logic and the memory banks.
reasoning
Software and its engineering
train_126625
Similarly, the introduction of a networking abstraction such as a low power listening MAC introduces a rather unpredictable use of the CPU and the radio device, possibly delaying other tasks.
the developer does not have complete control over the application, which makes it harder to ensure a bug-free program.
reasoning
Software and its engineering
train_126626
A general­purpose VMcould embrace any workloads including multimedia, while the hypervisor regards the VM as a resource allocation unit without any knowledge about the tasks running on the VM.
it is hard for the hypervisor to decide how much CPU is allocated to a VM even when a user knows the ex­act demands for the desired multimedia quality.
reasoning
Software and its engineering
train_126627
and ) adhere to SQL terminology.
or can be used as well within dynamic .nders.
reasoning
Software and its engineering
train_126628
Such optimizations are applied (in some cases) to enable vectorization.
the dynamic loop count of a run with and without vectorization are di.erent (refer to Table 8).
reasoning
Software and its engineering
train_126629
creating business document models, based on feature models it is necessary to perform a mapping between feature models and Core Components.
we discuss the mapping between CBFM and Core Components.
reasoning
Software and its engineering
train_126630
what conditions are being changed (variables, context, materials, etc.).
these three axes must be included in any set of terms used to characterize replications.
reasoning
Software and its engineering
train_126631
This broadcast is essential due to inherent nature of paralleliza tion of the programs, for example, since only one process is reading the data; all other processes must wait until they receive their data portions.
the communication overhead dominated the running time in multi-computers and the gain due to parallelization was limited.
reasoning
Software and its engineering
train_126632
However, deep expects input in a special .gen format.
we con­verted points-to constraints from LLVM IR into .gen format using a simple LLVM pass.
reasoning
Software and its engineering
train_126633
On the contrary, in case of the VF mode, the guest buffer needs to be mapped in Domain0 even before the actual hardware re­ceive event occurs.
netfront driver is modi.ed to pro­vide grants more aggressively i,e,.
reasoning
Software and its engineering
train_126634
Due to the in­ability of user-layer components to control components in lower layers, execution times can widely vary.
such estimates are typically gross over-approximations, resulting in underutilization of computing resources.
reasoning
Software and its engineering
train_126635
To meet this need, we aim to provide a novel architecture mPat, which is uniquely built upon real-time multi-source feeds to abstract human mobility for real-world applica­tions.
for the .rst time, mPat is able to provide insights on not only where and when but also how (e.g., by bus) almost all urban residents move.
reasoning
Software and its engineering
train_126636
In general, a worker node cannot immediately forward all the new slot information to both its downstream worker nodes because the snowball e.ect would double the mes­sage size in each step.
each worker node ran­domly picks some information and keeps the remainder for its next message.
reasoning
Software and its engineering
train_126637
a thread, forwards a Start signal to its subtrees before starting its own computation, and forwards a Done signal to its parent node only after receiving all the subtrees Done signals and completing its own task.
the implicit barrier at the end of OpenMP work-sharing constructs takes a time corresponding to the worst-case propagation delay through the tree, which is logarithmic in the number of threads.
reasoning
Software and its engineering
train_126638
Figure 6: Consistency Agent Migration Time on Concurrent User Number Figure 7: Consistency Agent Migration Time on Entity Number Our experiments show that Consistency Agent migration does not affect the real-time interaction of the CVE users, and so do other types of agents.
our proposed grid-based mobile agent approach will improve the scalability of LCVE without compromising its performance.
reasoning
Software and its engineering
train_126639
Figure 2 exempli.es the concept.
a .rst strand indicates that a change of the Communicator whichis usedby the Tra­deComponent canleadtochanges in the performance which can be measured via the sensor PerfS at the interface PI.A second strand indicates that a change of the Communicator which is used by the TradeComponent canleadtochanges in the communication costs which can be measured via the sen­sor CostS at the interface CostI.the controller ComS­elect is able to manipulate performance and communication costs by exchanging the Communicator which is used by the TradeComponent.
reasoning
Software and its engineering
train_126640
We have observed that the Internet delay space displays special structural and statistical characteristics in that it ad­heres to a power-law that extends over a signi.cant range of Internet distances, namely the intra-continental distances.
its dimensionality can be characterized using frac­tal measures, sensitive to non-linear as well as linear struc­tures in the delay space.
reasoning
Software and its engineering
train_126641
Second, because we allocate transactional data on a CDU boundary, no two allocations can share a CDU.
increasing the CDU size does not increase the number of con.icts for this benchmark.
reasoning
Software and its engineering
train_126642
Therefore, it is necessary to go back and run the authUML again to preserve the consistency, complete and conflict-free accesses.
our framework is flex- ible enough that allows changes in access control specifications.
reasoning
Software and its engineering
train_126643
Moving more than one page Whenatrapoccurs,itimposessomeoverhead.Thus,itmakessensetomovemorethanonepageineachtrap.Wetypicallymovedeightpagespertrap.Thisopti­mizationalsoreducesthespaceoverheadofthestatus-table andthe.rst-object table.
double mapping to-virtual-space in the beginning Thetraprou­tineneedstotouchaprotectedpagewithoutliftingtheprotection.asecondvirtualmappingoftheheapisrequired,whichisnotprotectedandisavailableforusebythetraproutine.Toreducethetraptime,weDoubleMaptheentireto-virtual-space toathirdvirtualspacewhencompactionbegins.Ofcourse,thismappingcanberunconcurrentlywiththeprogramrun.
reasoning
Software and its engineering
train_126644
The following three file systems are compared against hFS: (1) FFS: Although the original design of FFS was done more than a decade ago, it has received substantial improvements over the years [20].
fFS is a formidable file system to compare against.
reasoning
Software and its engineering
train_126645
In contrast, Molhado [8] is an SCM infrastructure that creates unique, persistent IDs for each entity, and treats the name as an attribute that can change.
it can track the history of an entity in spite of it being refactored.
reasoning
Software and its engineering
train_126646
However, access to the system mem­ory is initiated by the debug port hardware component and not through software instructions executed by the proces­sor.
the timings of the application code are not changed by the JTAG access.
reasoning
Software and its engineering
train_126647
n .4. ii.1 The reason is that cycle 1 and cycle 2 cannot provide protection for same span, cycle 2 and 3 can t either, so cycle 2,3 and cycle 1,2 can share spare units on span 3-5.
we can anticipate that the shared-p-cycles design can achieve better efficiency than the conventional p-cycles design.
reasoning
Software and its engineering
train_126648
Note that F(pj) jj S(pj)+v(t)+>v(t).
only the .rst newly j; ; arrived request for fuses formula (23); subsequent requests use formula (24) until fbecomes inactive.
reasoning
Software and its engineering
train_126649
We denote the set of generators at, a., u., v., w. de.ned above by C. By the product formulas of Table 2, every element of C is a linear combination of 2-long products from A and, conversely, every element of A is in (C).
in order to prove that ((A)) is 2-closed and to give a full description of the algebra operations, it is enough to show that the product of any two elements of C remains in (C)and to compute the product and scalar product of any two elements of C. Step 3.
reasoning
Software and its engineering
train_126650
In case, the j-th log-likelihood ratio does have the absolute minimum value of all K inputs, the minimum of the K inputs except the j-th one equals the second-to-minimum value.
on the K input LLRs to a given check node, we take the minima of all numbers except one, where we first omit the first one, then the second one, then the third one, etcetera in a check node the set of minima of K values, except the j-th one, can be stored using only 3 values (a dataset) instead of K=30: The overall minimum value, The overall one-but-minimum value, The index j for which the minimum input value occurs.
reasoning
Software and its engineering
train_126651
The same applies for documents concerning patients and by that also the noti.cations.
sending information about critical results in an analysis over systems which are out of the scope of the hospital would imply a breach of secrecy.
reasoning
Software and its engineering
train_126652
However, building data context information could be expensive because (a) data accesses are highly frequent, and (b) access patterns among objects are complex.
it is desirable to establish a program paradigm that can collect such information efficiently.
reasoning
Software and its engineering
train_126653
To avoid this exponential blow-up, we impose a total order on abstract locations such that if ai and aj are aliases, they both point to a common location *ak such that ak is the least element in the alias partition class of ai and aj .
in the init heap rule of Figure 8, ai points to *ak where k = i under constraint: (*ai = *aj ) .
reasoning
Software and its engineering
train_126654
If the main document and the embedded object are produced by software from different vendors, it is gen­erally impossible for one to know about the metadata in the other.
the data must be purged separately from each object and from the main document.
reasoning
Software and its engineering
train_126655
It seems like a natural step forward to provide the same degree of .exibility for parameteri­zation of templates.
this can be seen as making the generic constraints of a template .rst class entities of the template, in the same way that classes and inter­faces are.
reasoning
Software and its engineering
train_126656
The macros in the documents ensure consistency with the artifacts stored in the repository.
the user activated macros automatically retrieve and update the artifacts in the database.
reasoning
Software and its engineering
train_126657
The proposed hardware deadlock detection algorithm makes use of parallelism and can handle multiple requests/grants, making the proposed algorithm faster than the Os~,(1) algorithm[4], which actually has Os,,,(m x n) latency in multiprocessor systems.
the SoC par- allel deadlock detection algorithm/hardware is faster than any previously reported algorithms.
reasoning
Software and its engineering
train_126658
This is, however, incon­sistent with the relation obtained by looking at column two in the FAS, i.e., the relation obtained by this column is s1,2 = 3s2,2/2.
the two vector sets are incompatible, which of course is possible (and also highly likely) since the values in the vector sets are obtained from (expert) opinions and experiences.
reasoning
Software and its engineering
train_126659
DSM s as used by Baldwin and Clark and in earlier work do not appear to model the environment in which a design is embedded.
we were unable to model the forces that drove the design changes that Parnas hypothesized for KWIC.
reasoning
Software and its engineering
train_126660
On the other hand, modeling is more than drawing, but using formal languages which are intended to be readable both for humans and machines in most cases.
the problem of wrong data formats has to be tackled from both ends: On the one hand, students must be taught that using diagramming tools is more than drawing on a sketchpad.
reasoning
Software and its engineering
train_126661
Whenwelook attheBGPinformationavailableinthe .ows, a very large number of pre.xes are announced from multi­ple ISP-X PoPs.
in many cases ISP-X selects an exit PoP from a list of PoPs where the destination pre­.x is announced and in general it prefers the PoP that is geographically closest to the entry PoP.
reasoning
Software and its engineering
train_126662
Geronimo for example has 177 Javascript .les, 1004 XML con.guration .les, 19 con.guration .les, and 105 image .les.
it is possible that we missed some co-change relations among non-Java based artifacts or between non-Java and Java-based artifacts.
reasoning
Software and its engineering
train_126663
The virtual camera has a .eld of view of 40 degrees and a 4/3 aspect ratio.
12 images are needed to build a complete panorama.
reasoning
Software and its engineering
train_126664
The di.erence is that in interval Ii, we assume device Di is transitioned while in interval Ii-1 we assume Di is not transitioned.
this totally covers at most 2mcases were Ropt = d-Bi,i=1 ...m.Fur­ther, we also evaluate the energy consumption when Ropt = c and Ropt = d. an implication of Corollary 1 is that there are at most 2m+ 2 cases that need to be examined to determine the optimal solution.
reasoning
Software and its engineering
train_126665
A, and is the empty set otherwise.
on the range (1.5, 2), we have µout((1.5, 2)) = µ((0.5, 1.0) n [0.5, 8)) + µ(Ø n (-8, 0.5)) = µ((0.5, 1.0)) = 0.05 and on the range (9, 12), we have µout((9, 12)) = µ((8, 11) n [0.5, 8)) + µ(Rn (-8, 0.5)) = µ((8, 11)) + µ((-8, 0.5)) = 0.25 i.e.
reasoning
Software and its engineering
train_126666
To move to the root of the tree hit home to move to the last item hit end."
the use of a standard control allows the user to access meaningful information provided by JAWS.
reasoning
Software and its engineering
train_126667
Since the first eight bits extracted from the final codestream whose decimal value equals 255, which is an un-embeddable index value, this means that it is transformed from an embeddable index value with secret digits embedded.
the decoding procedure falls into Case 2, so the original index value 0 is reversed, and the base-4 secret digit 0 is extracted by mapping Figure 4.
reasoning
Software and its engineering
train_126668
A process can be a simple process, including only a sequence of activities, or a complex process including a controlled composition of services and activities performed under condition, in parallel, or encapsulated in sub processes.
to define a process, it is necessary to define: activities the basic elements of a process.
reasoning
Software and its engineering
train_126669
Early completion may eventually cause the CPU to be idle, which, in turn, results in energy waste since the device still consumes energy during the idle time.
there is a potential to save more energy by avoiding or reducing the idle time.
reasoning
Software and its engineering
train_126670
If there are more than one point on the time axis that correspond to .i × Ui max, we choose the latest point.
ti is probabilistically assured to accrue at least the utility percentage = Ui(Di) Umax .i i , with probability .i.
reasoning
Software and its engineering
train_126671
Our meta-tools are used to generate complex visual modeling tools, and these modeling tools could benefit from the addition of various critics.
we wanted to extend our Marama meta-tools by embedding a critic design and generation component.
reasoning
Software and its engineering
train_126672
In frameworks such as EMF, this separation is not sup­ported.
the model conformity is validated against the same metamodel combining the two dimensions.
reasoning
Software and its engineering
train_126673
By using the proposed description rules, it is possible to relate I/O message elements of a system constructed using CORBA to the message parameters in OWL-S.
we can apply OWL-S to CORBA with our description rules [3].
reasoning
Software and its engineering
train_126674
Thus, the potential never increases, and it decreases by at least 1 after every deductible quantum.
the number of deductible quanta is at most 2T8 +log.
reasoning
Software and its engineering
train_126675
In intentional programming, complete programming language are imple- mented as a set of intentions, whereas ELIDE oniy supports simple modifier definitions for Java.
an in- tentional abstract syntax tree is modelled at a fine level of detail, whereas in ELIDE it is coarse-grained.
reasoning
Software and its engineering
train_126676
It is also of questionable worth since we expect program­mers will often .x one error and recompile.
we extend our algorithm to recursively search for good error messages in one subtree while simultaneously removing one or more other problem­atic subtrees.
reasoning
Software and its engineering
train_126677
In that case, DMGexit (cb)= N, a condition that is checked by the out new transfer function given in .gure 11.
the transfer function depends only upon DMGexit .
reasoning
Software and its engineering
train_126678
The major shortcoming of static schedulers lies in the fact that they assemble a schedule prior to work.ow execution, which is then strictly abided to.
changes in the runtime environment make even elaborate static schedules suboptimal.
reasoning
Software and its engineering
train_126679
[1] Figure 4: Playlist aspect When implementing CreditsAspect, the programmer as­sumes a one-to-one relation between the playing of a song and its purchase.
the compositional intentions for its advice are the following: Ignored: NoProceed() After: MultiProceed() These compositional intentions mean that the advice is assumed to not be executed if the join point is ignored by another advice (.rst line), and for each execution of the join point the advice is assumed to be executed as well (second line).
reasoning
Software and its engineering
train_126680
In other words, there must exist a function F ' that connects the value of v with the value of v'; i.e., v = F (v).
we consider the following types of data dependency: 1.
reasoning
Software and its engineering
train_126681
We have: if execution level is in the model then one of { task , ISR } is in the model; if ISR is in the model then one of { ISR1 , ISR2 } is in the model.
if execution level is in the model then one of { task , ISR1, ISR2 } is in the model.
reasoning
Software and its engineering
train_126682
Each feature is given a cardinality, even mandatory and optional features, which are represented as cardinality [1..1] and [0..1] respectively.
a cardinality cannot be removed or added but only updated which can result in range inconsistencies.
reasoning
Software and its engineering
train_126683
A recon.guration transaction should take a software from an initial state to a .nal and consistent state.
a recon.guration transaction can be committed only if the resulting system is consistent, i.e., if all integrity constraints on the .nal system are satis.ed.
reasoning
Software and its engineering
train_126684
However despite the recorded gains, the statistical (paired t-test) p values obtained for the Control and Experimental groups were both greater than the required value of 0.05.
we could conclude that the mean selection score for both groups are not statistically different.
reasoning
Software and its engineering
train_126685
In our case, where there are only two vot­ers, the set of solutions degenerates to include the two input permutations, as well as any intermediate permutation in the bubble-sort path between them.
the Kendall distance is not useful for our problem.
reasoning
Software and its engineering
train_126686
After exit­ing a nested query, it may be necessary to recover the outer query s insertion points, so that code for subsequent operators may be in­serted.
we de.ne the stack, S, as containing (ai,µj ,.k) triples, where i, j, k = 1 are the nesting levels of each insertion point.
reasoning
Software and its engineering
train_126687
Especially, for safety consideration, for a train under CTCS-3, RBC can only grant it the MAs before a CO segment, and the train needs to ask the permission of the driver before moving into the CO segment.
both the speed limits for CO segments are initialized to be 0 in CTCS-3.
reasoning
Software and its engineering
train_126688
An object decides on subscription and APUM frequency based only on local information (derived from PUMs and APUMs).
it may be possible for an object, say Oi, to subscribe to another object s, say Oj s, local channel without Oj subscribing to Oi s local channel.
reasoning
Software and its engineering
train_126689
The real world auction trace that was considered does not re.ect a high intensity of intervals as may be expected for scenarios such as the online distributed tra.c navigational data broadcasting.
we used synthetic trace data to generate a wide range of problem instances that could re.ect various parameters of di.erent applications.
reasoning
Software and its engineering
train_126690
However, unlike the communicability evaluation, semiotic inspection does not involve observing users interaction with systems.
the two methods capture differ­ent communicability aspects and perspectives.
reasoning
Software and its engineering
train_126691
The corresponding transfer function does not .nd a loop-level kernel because SCC1 S (h1)represents the com­putation of the temporary variable h1 associated with the loop index.
the analysis of the structural use­def chain SCC1 A(a1...3) .
reasoning
Software and its engineering
train_126692
M2 is a schema mapping speci.ed by an st-SO dependency.
m '' is speci.ed by an st-SO dependency.
reasoning
Software and its engineering
train_126693
However, it differs from them in three aspects: i) lock acquisition and actual writes are shifted to commit time; ii) the validation proce­dure is modi.ed to satisfy the new transactional require­ments; and iii) the necessary information is saved in local semantic read-sets and write-sets.
oTB gains the following bene.ts over pessimistic boosting.
reasoning
Software and its engineering
train_126694
Finally, the Tightness Coe.cient Tij of nodes ui and uj isde.ned as the 1 http://wiki.limewire.org/index.php?title=Overview ratio of arcs that their neighbors have in common and the maximum numberoflinksbetweentheir neighbors; thegoal of this parameter is exploiting the high average value of the clustering coe.cientinthe a.nity network,in ordertogive priority to items related to most representative neighbors.
relevance reli(sk) is computed as a function of Tij, rate(uj,sk)and A.
reasoning
Software and its engineering
train_126695
Predictor functions .j are used to predict the property values pj resulting from the composition of individual component properties ci ci pj , where pj represents property pj of component ci in the com­position.
we have: ci1 si2 qpj = .j (qp j ,qp j ),j .
reasoning
Software and its engineering
train_126696
We controlled this learning effect by first performing the functional test case creation and in the second run the code inspection.
the students had no knowledge of the concrete implementation (i.e., the source code) or the defects included (as test execution was not performed by them) when they performed the inspection.
reasoning
Software and its engineering
train_126697
Since each sweep of u. costs 2. logN moves, the total number of moves is: h X N.(h +1, 1)2. log N. .=0 This quantity is the sum of the sweep costs at each level, until the virtual node needs its .rst rebalance.
the amortized number of element moves is h X 1 N.(h +1, 1)2. log N. (9) N .=0 PROOF OF THEOREM 3: We bound N.(.,1), the number of sweeps of the leftmost child u. at level .
reasoning
Software and its engineering
train_126698
We argue that this is mainly due to the lack of efficient and clear abstractions that developers can use to incorporate interactivity into their applications.
interaction represents a major overhead for developers, and users are faced with inconsistent interaction models across different displays.
reasoning
Software and its engineering
train_126699
We start by considering j1 and showing that aj1 = bsl i .
let us assume that j1 was assigned in phase p, for some p,1 = p = k. If aj1 exceeds the value b· sl i then the a-values for all the remaining l- 1 clients also exceed b· sl i by the timeAlgorithm3 entersphase p. the total payment from S(i)exceeds b· si.
reasoning
Software and its engineering