id
stringlengths
7
12
sentence1
stringlengths
5
1.44k
sentence2
stringlengths
6
2.06k
label
stringclasses
4 values
domain
stringclasses
5 values
train_126800
Fortunately, executing a change object results in an undo object being computed, since change ob jects im­plement the command pattern as described in the gang of four book [3].
when we move right in Fig.
reasoning
Software and its engineering
train_126801
Moreover, the paper establishes that t< n/2and z = k are two tight bounds of the k-set agree­ment problem, when considering the (Oz)1=z=n family of failure detectors in an asynchronous message-passing system (Theorem 1).
among all the classes described in Figure 1, Ok is the weakest class for solving asynchronous k-set agreement (hence, the algorithm is optimal in that re­spect).
reasoning
Software and its engineering
train_126802
Furthermore, most dependence tests assume a.ne subscript expressions.
our method is restricted to a.ne functions of the loop indices.
reasoning
Software and its engineering
train_126803
Hence, at most |W |= sds = s/8e of the lookup requests fail.
at most O(log n) attempts are necessary until all requests have been served.
reasoning
Software and its engineering
train_126804
This is basically means that a complex function that has too many variables and parameters can be decomposed into some sub-functions known as constituents (parts) which are inter-related but have less variables compare to the original problem.
the original function can be reconstructed from those constituents.
reasoning
Software and its engineering
train_126805
No matter which mapping is used to perform the write, if the code-page content is ultimately changed, the code cache must be updated accordingly.
to ensure correct­ness of the code cache coherence solution, we must monitor not only that particular page-mapping, but also all present and future page-mappings that map on to that physical page.
reasoning
Software and its engineering
train_126806
Two obstacles were given a severity rating of 1, five were given a rating of 2, six were given a rating of 3, and 45 obstacles were given a severity rating of 4.
applying this relatively conservative rating scale, the vast majority of obstacles introduced a minor delay into the process of carrying out the assigned tasks.
reasoning
Software and its engineering
train_126807
However, having too many layers will result in increasing the overhead due to the added interface and communication complexity between the layers.
it is important to design architecture with an optimal number of layers for sensor web applications [7].
reasoning
Software and its engineering
train_126808
2.2 Dynamic Repartitioning The repartitioning strategy described in Section 1, whereby dur­ing a quiet time servers are taken of.ine to be moved into new clusters, has several problems.
first, reducing electricity use requires running fewer servers at relatively high utilization levels rather than more servers at lower utilization.1 reducing the capacity of the network to repar­tition is dif.cult while sustaining the required query throughput.
reasoning
Software and its engineering
train_126809
The OOG extraction nondeterministically selects a label for a given object o based on the name or the type of one of the references in the program that points to o.
detecting renames ensures a developer can still rename .elds or local variables or types without impacting conformance.
reasoning
Software and its engineering
train_126810
While PTD refers to a sub-optimal structure of a system, ETD refers to PTD, which in addition has a negative impact on future development.
while PTD is the source of ETD and hence is important to identify, only ETD must be e.ectively dealt with.
reasoning
Software and its engineering
train_126811
Dependencies cannot be managed at an individual level only at the component level.
the resulting compositions may be slightly larger than necessary.
reasoning
Software and its engineering
train_126812
A single modification to the animations during run-time results in a stop of the animation viewer and regeneration of a new scene file.
the real­time approach has the advantage over the offline approach and is adopted in our framework.
reasoning
Software and its engineering
train_126813
Directoryentriesarestored as .lenameandinodenumberpairs, and arepacked intopage size chunks.
file names are restricted so that adirectory entry always .tsinto256bytes.achunk with a hole can always contain another directory entry.
reasoning
Software and its engineering
train_126814
When imgin and imgout are mapped in the same bank, an access to imgout is only scheduled after K2accesses to imgin.
the frequency at which accesses to both data structures interfere is much lower than for kernel and imgin.
reasoning
Software and its engineering
train_126815
4.2 Peer Discovery Building the application s communication on top of point-to-point unicast connections means that a sender has to contact each receiver individually.
a peer needs to know the IP addresses of all the peers it is collaborating with.
reasoning
Software and its engineering
train_126816
Thus, the apache package should only be installed on a system that does not also have the apache-modules package, any instance of the jserv package with version less then 1.1.3 and so on.
to install apache, the package manager must .nd out which other packages must be installed such that ulti­mately the system contains a set of packages that meet all the requirements speci.ed in the distribution metadata .le.
reasoning
Software and its engineering
train_126817
Lazy restore can take quite some time to completely .nish because memory is restored only when it is accessed by a VM or touched by a slow-running background thread.
the performance is bounded by the memory size of the saved VM and the speed at which the background thread retrieves memory.
reasoning
Software and its engineering
train_126818
A change in the API could affect some or all external applications connected in this API.
this change will force external developers to adjust and retest their applications to work on the platform.
reasoning
Software and its engineering
train_126819
they can be checkpointed/restarted with the appropriately smaller sizes.
the framework becomes com­pletely self-tuning.
reasoning
Software and its engineering
train_126820
Now we assume node G updates its replica of the file and informs the other nodes by sending an in­validation packet over the validation mesh.
all other nodes except G remove their replicas of the file from their storage and the validation mesh is torn down.
reasoning
Software and its engineering
train_126821
During this study, we plan to test the co-evolution approach in projects with an existing code base and known architecture.
at .rst, we use our ap­proach to generate a PCM from the legacy source code.
reasoning
Software and its engineering
train_126822
Second, based on having explicit vari­ability models, business partners and SDOs may de.ne busi­ness document models through con.guring the variability model.
variability model con.gurations serve as trace models between the generic Core Components and derived business document models.
reasoning
Software and its engineering
train_126823
A .ow graph is constructed so that if declaration A in.uences declaration B, then the graph will contain a path from A to B.
our global in.uence relation is just the transitive closure of primitive in.uences (directed edges in the .ow graph) induced by the program text.
reasoning
Software and its engineering
train_126824
Second, leaks have the potential to impact not only the application that leaks memory, but also every other application run­ning on the same system; because the overall amount of memory is limited, as the memory usage of a leaking program increases, less memory is available to other running applications.
the performance and correctness of every running application can be impacted by a program that leaks memory.
reasoning
Software and its engineering
train_126825
The original cumulative constraint provided by MiniZinc [15] could not be used because it was not able to handle the two different task types present in MapReduce jobs: map tasks and reduce tasks.
a modified cumulative constraint, called mycumulative, is developed that ensures that map tasks and reduce tasks are only scheduled on the map slots and reduce slots of the resources, respectively, and also the capacities of the resources are not violated.
reasoning
Software and its engineering
train_126826
For a component to be an independent unit of compiled code, it must be completely separate from the environment in which it is de.ned and compiled precluding any explicit references to other components.
a component com­pletely encapsulates its features and can only be deployed in its entirety.
reasoning
Software and its engineering
train_126827
It should be noted that all data in a block is fetched at cache-misses and kept coherent as a unit.
it is important whether multiple shared accesses are issued onto data in the same block.
reasoning
Software and its engineering
train_126828
Since the subsumption rules give rise to constraints between types, we cannot give a coherent constraint solving strategy that is still complete we either need to reject certain reasonable programs or we need to solve such constraints too aggressively leading to an incomplete inference algorithm.
our implementation uses the simple strategy since neither lifting nor subsumption provide a satisfactory improvement.
reasoning
Software and its engineering
train_126829
We found a Naïve Bayesian model that uses Markov, Hour and Cell-Tower (the blue plots in Figure 8) gives good overall accuracy across the three behaviors, which has also shown effective in previous work (e.g., [20]).
we compared the performance of this Naïve Bayesian model with Reflection.
reasoning
Software and its engineering
train_126830
Second, when a failure is detected by NTFD, NTFD will send the failure notification messages to the registered processes on the remote hosts.
the message cost only increases linearly with the number of failed processes.
reasoning
Software and its engineering
train_126831
This forces A to make its decisions without knowing how much time is left.
a has to attempt to discover new bugs as early as possible.
reasoning
Software and its engineering
train_126832
This would enable practitioners to fully exploit the benefits of SRA adoption and usage.
our research goal is: Which are the artifacts that compose an SRA and how such arti­facts are designed, reused, and used?
reasoning
Software and its engineering
train_126833
Our approach also deals with a graph that contains the operations as nodes and the relations between nodes as edges, similar to the serialisation graph.
our approach based on a graph and constraints between operations can be mapped to the ACF framework that represents a theoretical formalism for modeling replication.
reasoning
Software and its engineering
train_126834
A data race is a pair of accesses to the same variable by different threads, where at least one access is a write and the accesses are not ordered by synchronization.
if an execution has no data races, the variable (e.g., x in Figure 1) is interference-free for the duration of the execution of the region.
reasoning
Software and its engineering
train_126835
We assume that there are no transitions leading from and to the states representing regions where obstacles are present.
each trajectory of the robot in the said subset of the environment is obstacle-free.
reasoning
Software and its engineering
train_126836
Also, at any point in the interleaving, if the last step of S(i) that has been performed up to that point is sk, then we de.ne the cache state at that point to be Ck.
the cache state only depends on which steps of S(i) have been performed, not on how S(i) and T ( j) are interleaved.
reasoning
Software and its engineering
train_126837
Therefore, given that the magnitude of RRmn(t) is r at time t, the probability that node n is located within zone 1...i after t seconds is given by 11 r = Ri - d0 P {d(t)= Ri|Rmn(t)= r} = 2 22 1 -1 d0+Ri -r cos r>Ri - d0 p 2d0Ri (10) From Equation 9, the probability density of r is given by 2 r f (r, t)= 2r e - am+an .
am+an JRi+d0 P {d(t) = Ri} = P {d(t) = Ri|Rmn(t)= r}f (r, t)dr r=0 (11) According to [18], the node distribution is similar whether the link has to be continuously available or not.
reasoning
Software and its engineering
train_126838
The loss of output-data leads to wasted user time allocation, which is very precious and obtained through rigorous peer-review.
a timely o.oad can help optimize both center as well as user resources The need for timely data o.oading is also fueled by the, often, distributed nature of computing services and users job work.ow, which implies that data needs to be shipped to where it is needed.
reasoning
Software and its engineering
train_126839
Station D contains a database containing title and pay.
to compute the above query mobile agent will have to go to PEs where databases are distributed and perform necessary database operations.
reasoning
Software and its engineering
train_126840
Regarding SQL statements, the Stones library provides three different features.
when an SQL statement was created without the Stones framework, we removed it from the usage count of the feature that would have been the most appropriate.
reasoning
Software and its engineering
train_126841
For the size of each table en­try of path bf, rule (B ) incurs O(#edge.2) due to the .rst hypothesis, and rule (L ) incurs O(#edge.2) due to the sec­ond hypothesis.
the total size of each table entry is O(#edge.2).
reasoning
Software and its engineering
train_126842
For instance, in Chrome milestones critical defects are being .xed in less time than the Firefox project.
the feedback generated from previous knowl­edge can help testers and developers involved in .xing the security defect decide the severity of the security bug.
reasoning
Software and its engineering
train_126843
When the overhead for turning on/off is non-negligible, it will be not worthwhile to turn off the processor if the idle interval is not long enough.
procrastination algorithms such as those in [5, 9, 10] resolve this problem by aggre­gating idle intervals to reduce the number of mode switches and the overhead.
reasoning
Software and its engineering
train_126844
First of all, it gives designers the freedom to express their ideas freely, either using handmade sketches or tool-oriented wireframes.
they do not need to rely only on the tool s own environment.
reasoning
Software and its engineering
train_126845
We observed many calls to the sifting algorithm during the BDD building process each taking several minutes to complete.
we do not see any real bene.ts of using sifting to build BDDs for feature models.
reasoning
Software and its engineering
train_126846
One concern in this regardis that adding new atoms must notbreak the existing index-to-atom mappings that may be in use.
thedictionary is append-only.
reasoning
Software and its engineering
train_126847
In j=1 p a (L k-1 j=1 p this case, Copt = Cswitch + Sj)+: and Ctotal = k · total Rtr L k ( j=1 Spj) Copt Cswitch + .
we have Ctotal >a · total.
reasoning
Software and its engineering
train_126848
Interceptors adapt re­quests, replies and exceptions passing through the ORB.
the generic interceptors do not modify the component s behaviour.
reasoning
Software and its engineering
train_126849
There are two approaches how the type information can be provided by the JVM: Actualising during run-time: The type information of the memory cells is actualised on every execution of a Java instruction.
when a migration is requested the type information of the memory cells is provided.
reasoning
Software and its engineering
train_126850
The memoized version mf captures this by .rst checking if x is positive or not and then exposing the underlying value of y or z accordingly.
the result will depend on the sign of x and on either y or z .
reasoning
Software and its engineering
train_126851
Similarly, while analyzing a polymorphic type, the term-level con­struct typecase must bind the kind of the quanti.ed type variable.
we introduce kind abstraction (.
reasoning
Software and its engineering
train_126852
In particular, Dimension must be able to decode both the source and target binaries, and be able to encode the target one.
we need a new version of Dimension when the format of either the source or the target binary changes.
reasoning
Software and its engineering
train_126853
A .ne customizability level requires smart users be­cause some details about the underlying infrastructure can­not be hidden anymore.
openNebula is more suited for private cloud where the entire environment is trusted and users have more skills.
reasoning
Software and its engineering
train_126854
If visitor #3 goes back and visitor #2 becomes reachable again, a new instance of the video service has to be deployed on visitor #2 s host.
if visitor #3 is near from video server and if it has enough energy, it is not necessary to use the visitor #2 as a relay and the platform can directly deploy the assembly and send a black &#38; white video.
reasoning
Software and its engineering
train_126855
Indeed, the right granularity/precision of tracking varies, often within the same application [2, 19].
it appears to be a good idea to leave the choice of dependence checking mechanisms to the client (programmer and smart compiler) of the SMTX sys­tem.
reasoning
Software and its engineering
train_126856
To avoid double counting we use a polynomial algorithm to find the maximal partition of coallocatable objects that come from a coallocation set.
in Figure 2(a) a weighted miss reduction for coallocation set {A, E} only accounts for coallocating e6, a12, a14.
reasoning
Software and its engineering
train_126857
INTRODUCTION In the permanent ongoing production of Saarstahl AG, a variable set of uncertainties have a strong influence on the current planning.
flexibility and fast reorganization after such interferences at any point is needed to ensure a production without interruptions which still meets the needs of the customers at acceptable costs.
reasoning
Software and its engineering
train_126858
In order to enable a transpar­ent nesting of patterns and services, we need to make sure that these interfaces match.
work.ow patterns also implement a start method.
reasoning
Software and its engineering
train_126859
Thus there could be a combi­natorial number of mined LSCs.
it would be better to mine a representative set of LSCs.
reasoning
Software and its engineering
train_126860
For larger software designs, formal veri.cation using model checker often suffers from the state space explosion problem.
we need to use ab­straction techniques in order to alleviate the bur­den for the back-end model checker.
reasoning
Software and its engineering
train_126861
The .nal property is that all objects in both models are identi.able.
each object has a locally unique identi.er, which is unique in the model that the object resides in.
reasoning
Software and its engineering
train_126862
The input and output streams of the adjacent Kernel1 through Kernel6 are conformable.
these six kernels can be fused to increase both the granularity and locality.
reasoning
Software and its engineering
train_126863
In order to attach reactive behaviors to context changes, in addition to the method isActive, a context has to declare the events indicating context changes in its interface.
for abstracting from concrete contexts we need a means to abstract from concrete events.
reasoning
Software and its engineering
train_126864
DCR: We examine how dynamic changes in the com­putational infrastructure a.ect work.ow scheduling.
the RSD parameters responsible for long­term changes in the performance of a VM was varied between 0, 0.125, 0.25, 0.375, and 0.5.
reasoning
Software and its engineering
train_126865
This knowledge helps to significantly reduce maintenance efforts for variability handling of Simulink models.
it helps to increase the maturity of variant-rich models.
reasoning
Software and its engineering
train_126866
This di­versity can be highly complex as shown in the example in Figure 4.
we additionally proposed calculation algorithms and a possible approach for a largely automated computation of variants.
reasoning
Software and its engineering
train_126867
The code is generated by traversing the feasible .ring schedule found, and detecting the time where the tasks are to be executed.
the generated code can be seen as a cyclic executive, where the tasks are executed in accordance with previously computed schedule.
reasoning
Software and its engineering
train_126868
Then, for processes containing hiding and parallel composition, we show that using axioms as rewrite rules, one can eliminate these op­erators.
for each two PEPA processes p and q such '' that p .
reasoning
Software and its engineering
train_126869
As Table 1 shows, this is indeed the case even in our MI hierarchies.
a type eis allocated with an array rof size l e,v e,llb(eau taan c,e,sft oar stb(eau,t (in SI, l e,vee,llb(eau)==tatn c e,sft otr s/b(eau,t), with entries for each bbc aan c,e sft otr s/b(eaucb Speci.cally, bbsoeis stored in location l e,vee,llb(b{uin array r. checking whetherbas+ecan be carried out by checking whether b indeed occurs in location l e,vee,llb(b,uof array r. The encoding is optimized by not storing bitself in this location, but rather an id, which is unique among all types in its level.
reasoning
Software and its engineering
train_126870
Defining an annotation for feature dependencies As figure 2 shows, an annotation declares which elements it tar­gets (A in the example).
this element (and all its subtypes) appears to possess the additional child r declared by the annotation.
reasoning
Software and its engineering
train_126871
But, C2 ensures that if such a situation were to occur, that C would have an overriding de.nition.
the .rst dispatch rule would apply and superclasses would not be considered.
reasoning
Software and its engineering
train_126872
1), since the previous call pro­ vided the question object containing a qid needed for activating a check_question link.
the exploring of the web service by the generic state machine is driven by acquiring more dynamic links.
reasoning
Software and its engineering
train_126873
For on­tologies, the situation is similar, except that ontology lan­guages and their semantics are, as a rule, alien to develop­ers.
the tools have to go much further in organizing and presenting the information supplied by the reasoner and existing in the ontology.
reasoning
Software and its engineering
train_126874
Reliability it must always be possible to .nd needed components regardless of current location and system state.
single points of failure are unacceptable in any design.
reasoning
Software and its engineering
train_126875
B enchm arks Figure 8: Energy, Performance and Code Size comparison using SO vs. SA algorithm trend of energy consumption values remains the same for all the benchmarks.
a comparison of the two allocation al­gorithms is presented for a smaller and .ner range of SPM sizes.
reasoning
Software and its engineering
train_126876
For each specific problem, an optimal configuration can be chosen as a solution.
the chain-of-configurators architectural style enhances the flexibility and increases loose coupling compared to the other styles where the monitor and configurator were fixed.
reasoning
Software and its engineering
train_126877
The last chunk can move within the kth sub-heap.
our algorithm needs at most nine additional entries outside the sub-heaps.
reasoning
Software and its engineering
train_126878
To avoid false negatives would require a much more precise interprocedural analysis that would be very costly, and itself might introduce additional false pos­itives.
we chose to implement an analysis of practical cost, which identi.es, we believe, the bulk of the e-c chains of interest.
reasoning
Software and its engineering
train_126879
Finally, it classi.es the potential faults that become executable by applying our mutation operators as delay­dependent potential faults.
jSPreventer provides developers with executable evidences of the delay-dependent potential faults which are not-easily-executable faults be­cause they are not executed without subtle network delays.
reasoning
Software and its engineering
train_126880
This is obvious by construction, since the nature of immutability is such that the only writes are initializing writes.
there cannot be a write after a read (since this would imply a read of uninitialized data), nor can there be a write after a write (since this would imply mutation).
reasoning
Software and its engineering
train_126881
6http://www.eclipse.org/mylyn/, veri.ed 15/06/11 7http://www.eclipse.org/, veri.ed 15/06/11 3.2 Parsing Social NewsWeb Sites The intention of the second component is to characterize recent posts on social news web sites, such as digg, reddit and mashable.
for new article links8 posted on any of a user-de.ned set of social news web sites, our ap­proach follows the link to the referenced article and charac­terizes the article using a simple heuristic that abstracts an article to the technical terms that occur the most frequent in each article.
reasoning
Software and its engineering
train_126882
The inner loop has a tight dependence chain as the XVector element written to in one iteration is read in every subsequent iteration.
the only possible way to parallelize this loop is to violate sequential semantics by not enforcing some of the true dependences.
reasoning
Software and its engineering
train_126883
fAllowedW checks if a process has exclusive access to a .le.
if we were to write defensive code, before attempting .le access we would .rst check that our process had access to that .le.
reasoning
Software and its engineering
train_126884
Due to noise in the mobile sensors, the dead-reckoned trajectories are accurate in the beginning, but diverge from the truth over time.
outdoor lo­calization schemes like CompAcc [2] have triggered periodic GPS measurements to recalibrate the user s location.
reasoning
Software and its engineering
train_126885
As a re­sult, the path numbering algorithm will not assign a num­ber to paths that terminate at the back edge.
the instru­mentation on cold back edges ignores the value of the path register and always records an execution of a cold path2 .
reasoning
Software and its engineering
train_126886
Law enforcement -- If we understand how multiple modalities combine during expressive behavior, improved methods for detecting lies behavior may be developed.
subtle inconsistencies between modalities may prove to be an important indicator of deceptive behavior.
reasoning
Software and its engineering
train_126887
An awareness computer needs to only run software that mirrors the display of the primary computer.
it does not need to access proprietary applications or access proprietary data.
reasoning
Software and its engineering
train_126888
This mechanism allows the resource contributors to quantify the percentage of his/her contribution for anony­mous computing on Caravela network.
these capabili­ties of the worker and the broker servers implement a secure environment for GRID computing.
reasoning
Software and its engineering
train_126889
If a set of services is incompatible, the detailed measures and the mismatch list help to understand what parts of these services do not match.
the mismatches can be worked out using adaptation techniques, and service composition can be achieved in spite of existing mismatches.
reasoning
Software and its engineering
train_126890
Each combinator associates a particular type shape with an interference pattern.
a composition that does not meet the type shape required by the combinator fails to type-check.
reasoning
Software and its engineering
train_126891
This mapping is not exact, however, because neither the language nor TM system restrict exactly what memory can be passed into a method of an open atomic class, and the TM system performs a vanilla open­nested commit for a nested transaction, not a safe-nested commit.
it is unclear what exact guarantees are provided with respect to serializability and/or deadlock freedom.
reasoning
Software and its engineering
train_126892
After Q Until P, which is concerned with the event sequence after an occurrence of the starting delimiter, Q, and before an occurrence of the ending delimiter, P. P is not required to occur for this scope to exist This scope may occur repeatedly over the course of the system execution.
with the restrictions noted above, there is limited interaction between the intent and the scope these restrictions allow us 17 to propose a rather straightforward way to apply scopes to property pattern templates.
reasoning
Software and its engineering
train_126893
The length of the preamble should be long enough for synchro­nization between the AP and all UWB devices in the net­work, i.e., the preamble time Ta of i-th device is determined by di.
tf rame and TAC K are obtained as Tf rame = Ta +(pl +RUIh)/Ri +(PHYh +MACh +HCS +FCS)/R0 and TAC K =Ta +(PHYh +MACh +HCS)/R0.
reasoning
Software and its engineering
train_126894
We studied the feasibility of using these algorithms in order to categorize the right participants in the same group to be further analysed qualitatively.
two issues have been tested in this research: a) does this method (the use of clustering algorithms) of data summarization work for this kind of research (i.e.
reasoning
Software and its engineering
train_126895
The containers will have a label with the corresponding abstract graph.
some containers will be considered as idempotent containers when they are labeled with an idempotent abstract graph.
reasoning
Software and its engineering
train_126896
The coin toss decides whether to move to a node or not.
successive sampled nodes on the random walk may be between 0 and B hops apart based on the outcomes of these decisions.
reasoning
Software and its engineering
train_126897
However, since the students in Hbg have participated in a course where they developed a re­quirements document for a taxi management system, they only had a brief introduction.
the general introduction Table 1: Schedule for the experiment.
reasoning
Software and its engineering
train_126898
In addition, we want to be able to work with possibly large instances, to measure how ef.cient is the transforma­tion generated by a system.
we cannot rely on known tree-similarity measures, like, for example, edit distances [7], which are of high complexity and therefore would prove too expensive.
reasoning
Software and its engineering
train_126899
In this case, there is only one map block.
its shape is a single-level 3D shape of {[0 : M , 0 : N , 0 : K]}.
reasoning
Software and its engineering