What Is Fine And Coarse Granularity In Software Testing In Pdf

what is fine and coarse granularity in software testing in pdf

File Name: what is fine and coarse granularity in software testing in .zip
Size: 1226Kb
Published: 12.04.2021

Testing is a crucial part of the software engineering process.

While sensory processes are tuned to particular features, such as an object's specific location, color or orientation, visual working memory vWM is assumed to store information using representations, which generalize over a feature dimension.

Subscribe to RSS

While sensory processes are tuned to particular features, such as an object's specific location, color or orientation, visual working memory vWM is assumed to store information using representations, which generalize over a feature dimension. Additionally, current vWM models presume that different features or objects are stored independently. On the other hand, configurational effects, when observed, are supposed to mainly reflect encoding strategies.

We show that the location of the target, relative to the display center and boundaries, and overall memory load influenced recall precision, indicating that, like sensory processes, capacity limited vWM resources are spatially tuned. When recalling one of three memory items the target distance from the display center was overestimated, similar to the error when only one item was memorized, but its distance from the memory items' average position was underestimated, showing that not only individual memory items' position, but also the global configuration of the memory array may be stored.

Finally, presenting the non-target items at recall, consequently providing landmarks and configurational information, improved precision and accuracy of target recall. Similarly, when the non-target items were translated at recall, relative to their position in the initial display, a parallel displacement of the recalled target was observed.

These findings suggest that fine-grained spatial information in vWM is represented in local maps whose resolution varies with distance from landmarks, such as the display center, while coarse representations are used to store the memory array configuration. Both these representations are updated at the time of recall. This is an open-access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files. Competing interests: The authors have declared that no competing interests exist. Early students of cognition viewed the relation between memory and perception as analogous to that between a portrait and the scene portrayed [1] , [2]. Moreover, a long-standing intellectual tradition has since held that all memories are, or can be, spatially organized, since imposing a spatial structure facilitates the maintenance and recall of information, whether visual or conceptual [2] , [3].

While a prominent contemporary account of working memory has embraced this original metaphor of visual memory as a sketch of previously viewed scenes [4] , [5] , recent investigations examining the limits to the information that can be held in visual working memory vWM , do not support a spatially based, analogical model of vWM [6] — [9]. Initially, the observations that the ability to detect changes between subsequently presented scenes degraded rapidly when the scenes contained more than three or four objects, regardless of their complexity, led to the suggestion that visual data are stored in a limited number of object specific slots, each slot endowed with unlimited resolution [6].

Later, this model was revised to account for the fact that recall of visual data shows decrements whenever more than one object is held in vWM. The revised model suggested instead that slots have limited resolution and when the number of objects held in memory is less than the number of slots, more than one slot is used to store the same object [7]. Improved recall precision can then be achieved by averaging over independent memory representations.

An alternative interpretation of the gradual decline in recall precision with memory load is that limited resolution resources are used to represent specific visual dimensions, such as color, position and orientation [8] , [9]. Consequently, as the number of features in a given dimension increases, a smaller fraction of the global resource is available to represent each feature. This model predicts no upper limit on the number of features, and consequently objects, that can be held in memory, but shares with the former model the assumption that memory resources are not tuned to specific features within a given dimension.

These proposals imply that memory differs from sensory representations in visual cortex, which are tuned to specific features, such as the specific location, orientation or color [10] — [12].

A large body of neurophysiological work has indicated that during maintenance of information in vWM, sustained increases in neural activity take place in frontal and parietal areas, which are modulated by memory load [13] — [17].

The early slot model provided an elegant explanation of these findings, since the amplitude of the sustained neural activity appeared to track the number of slots utilized. However, neither the revised version of the slot model nor the resource model account for the effects of memory load on the amplitude of delay period neural activity, since both assume that memory utilizes all available slots or resources, irrespective of memory load.

Interestingly, more recent fMRI data suggest that visual information can be decoded from spatial patterns of BOLD activity in early visual cortical areas, during the delay phase of vWM tasks, even though no overall increase in BOLD activity is observed there [18] — [22]. Considering that these cortical regions contain neurons with receptive fields that span limited areas of the visual field, the aforementioned fMRI findings suggest that capacity limitations in recalling the details of a memorized scene depend on spatially curtailed processes and hence that a target's position may affect the resolution of its memory representation.

Moreover, neither slots nor resource models, which assume that features belonging to different objects are stored independently of each other, account for the finding that recall of a specific feature not only depends on the value of that feature, but also on the values of other features of the same dimension within the memory array [23] — [25]. Further evidence for global effects in vWM is provided by the finding that neural responses in parietal regions of non-human primates, performing a match to sample task, are affected by the spatial configuration of the memory array, but are invariant to the position of the array in the visual field [26] , [27] , suggesting that higher order neurons update their spatial selectivity, to gain access to the configuration of the visual scene.

We examined how precision and accuracy of spatial recall depends on local factors, namely the location of the memory target, global factors, namely the overall configuration of the items held in memory and configurational information presented at recall. We found that recall precision depends not only on the number of items held in memory, as previously reported [6] — [9] , but also on the target location, while recall accuracy depends on the overall spatial configuration of the memorized items.

Moreover, presenting configurational information at recall affected both the accuracy and precision of recall. We propose that spatial information is maintained in both local, variable resolution spatial maps, and coarse representations of the overall configuration of the memory items and that both representations are updated at the time of recall. To characterize spatial recall performance, the systematic and variable components of the recall errors were quantified separately.

The systematic error is the component that is consistently repeated over trials, while the variable error is the component whose value changes unpredictably trial by trial. We measured the precision in recalling the position of simple colored discs, when one or three were presented in the sample display Figure 1A.

Figures 1B and C show that the target location affected the standard deviation of the variable error of spatial recall. Specifically, targets located between the center and the boundaries of the display, were recalled less precisely than targets close to either the center or the boundaries of the display, suggesting that proximity to stable landmarks may facilitate the encoding and recall of spatial data in vWM. Moreover, the effect of target location on spatial recall was qualitatively similar whether the participants memorized one or three items.

However, memory load did change the overall recall precision, which was diminished when the participants had to remember three rather than one item. This result is important since it is at odds with the possibility that the effects of target location and memory load on recall arise at separate stages.

For example, a plausible hypothesis could have been that memory load effects reflect the limited capacity of working memory, while those of target location, the spatiotopic organization of early perceptual mechanisms.

However, if this were the case, the effects on recall variance of target location would be additive with those of memory load, which is contrary to the finding reported above. Figures 1D and E show, for each target location, the group averaged recall error's standard deviation, when memory load was three, as a function of the standard deviation of the error, when memory load was one.

The relation between the standard deviations is multiplicative rather than additive. We estimated, participant by participant, the best fitting additive and multiplicative models. Moreover, we found that the error standard deviation at each of the target locations was proportional to the square root of the memory load. In fact, the recall error standard deviation, when observers memorized three items, was 1.

These values are consistent with previous estimates of the effect of memory load on recall error [8]. These findings suggest that spatial WM depends on spatially curtailed representations, whose resolution scales with the overall memory load and the target location. A Participants memorized the location of either one or three items. After a pattern mask and blank interval, the target to be recalled was indicated by its color. B The standard deviation of the recall error is shown as a function of target azimuth, when the memory load is one in blue and three in red.

C The recall error as a function of target elevation. The variable error is smaller for targets closer to the center and boundaries of the display compared to intermediate positions.

D The standard deviation of the recall error when memory load was three is shown as a function of the error standard deviation when memory load was one for target azimuth, and E elevation.

Each point represents the group averaged error standard deviation at one of the nine target locations. The vertical and horizontal error bars are standard errors of the mean. The dash-dot line represents the group average best fitting multiplicative model. It is known that recall of spatial information from working memory shows systematic distortions, which depend on both stimulus and task related factors [28] — [30]. Some have also suggested that these biases reflect the reference frames used to encode spatial data in memory [31].

We characterized the spatial structure of systematic recall errors, separately for the two levels of memory load employed. Two patterns of systematic recall errors were found. Figure 2A shows that when the memory load was one, participants overestimated the target's distance from the display center, more prominently so along azimuth than elevation. Moreover, observers recalled the target at a lower elevation than its location in the sample display warranted.

However, when memory load was three, participants tended to underestimate the target's distance from the center of the screen Figure 2B. A Recalled targets were systematically displaced outward and downward in blue relative to their location in the sample display in black when the memory load was one.

B Recalled targets were displaced toward the center of the display when the memory load was three in red. C The six spatial components of the systematic error are shown, including constant offsets translation along azimuth and elevation, and four linear tensors. D Memory load only affected the divergence of the error field.

For the sake of convenience, the error size is expressed in degrees for the tensors as well. These values correspond to the displacement associated with each component, averaged over all target locations.

E Proportional recall bias in center of screen CS coordinates in blue when the memory load is one. F Proportional recall bias in CS in blue and center of the memory items' configuration CM coordinates in red when the memory load is three. Target azimuth, in CS coordinates, was overestimated both when the memory load was one and three. In addition, when the memory load was three, participants underestimated both target azimuth and elevation in CM coordinates.

To characterize the spatial structure of recall inaccuracies and gain further insight into the nature of the load effects, we reparametrized the systematic recall error using a set of four tensors, namely the divergence, rotation and two shear components of the vector error field Figure 2C. Next, we investigated why memory load affects spatial distortions in recall.

We observed that when participants had to keep three items in memory, the reported target location was shifted toward the locations occupied by the other two memory items, suggesting that spatial distortions, when the memory load increases, arise in a reference frame centered on the memory items.

Spatial distortions were thus modeled using two sets of linear regressors. The first set consisted of the target location in screen coordinates CS , the second of the target location in the center of the memory items' configuration coordinates CM.

The spatial configuration effects we observed may arise either because spatial data are smeared by vWM or because participants hedge their bets at recall, reporting an intermediate location, when they are not fully confident about which of the memory items is the recall target, and not because the location of the items held in memory is encoded in a reference frame centered on CM.

To examine this possibility, in experiment 2, the target was identified by presenting the non-target memory items at recall, but the non-target items were either translated away from the position they had occupied in the sample display 0. As shown in Figure 3B , the recalled target location was shifted on average by 0.

One possibility is that translating the non-target items shifts the origin of the reference frame, namely the CM, used to recall the target location. The other is that participants may have reported the target location, which preserves the distance of the target from the two non-target items. If the latter interpretation is correct, then displacing the non-target memory items' position by rotation at the time of recall should result in an identical rotation of the recalled target location.

Participants instead recalled the target location at a position rotated by 0. These findings suggest that participants did not simply memorize the relative distance between the items held in memory, but rather they encoded and recalled the position of the memory items in a reference frame centered on CM. This strategy is perhaps automatic, since participants were not informed that the non-target items' location at recall may be displaced. Moreover, enquiry after completing the experiment revealed that participants had failed to notice that the location of non-target items was occasionally changed at recall.

A The target was identified by displaying the position of the non-target memory items at recall. The position of the non-target items was either translated obliquely straight arrows or rotated around an axis through the display center curved arrows. B Translation of the non-target items, whose direction and magnitude is portrayed by a black arrow of normalized length, caused the recalled target location, portrayed by the red arrow, to be displaced in the same direction, albeit by a smaller magnitude.

C In contrast, following rotation, the recalled target location was displaced in a direction opposite the one required to preserve the distances between the memory items.

For illustrative purposes, the displacement of the non-target memory items is represented by the black line and the average displacement of the target items by the red line. To further examine the effect of target location and configuration of the memory items on recall, we changed the display's aspect ratio in experiment 3, thus modifying the distance between display boundaries and the target.

Semantic Approaches to Fine and Coarse-Grained Feature-Based Opinion Mining

Back to Search. Give Feedback. Description Conflict and dependency analysis CDA of graph transformation has been shown to be a versatile foundation for understanding interactions in many software engineering domains, including software analysis and design, model-driven engineering, and testing. In this paper, we propose a novel static CDA technique that is multi-granular in the sense that it can detect all conflicts and dependencies on multiple granularity levels. Specifically, we provide an efficient algorithm suite for computing binary, coarse-grained, and fine-grained conflicts and dependencies: Binary granularity indicates the presence or absence of conflicts and dependencies, coarse granularity focuses on root causes for conflicts and dependencies, and fine granularity shows each conflict and dependency in full detail. Doing so, we can address specific performance and usability requirements that we identified in a literature survey of CDA usage scenarios. In an experimental evaluation, our algorithm suite computes conflicts and dependencies rapidly.

Join Stack Overflow to learn, share knowledge, and build your career. Connect and share knowledge within a single location that is structured and easy to search. From Wikipedia granularity :. Granularity is the extent to which a system is broken down into small parts, either the system itself or its description or observation. It is the extent to which a larger entity is subdivided. For example, a yard broken into inches has finer granularity than a yard broken into feet. Coarse-grained systems consist of fewer, larger components than fine-grained systems; a coarse-grained description of a system regards large subcomponents while a fine-grained description regards smaller components of which the larger ones are composed.

It outlines the computers with multiple processing elements that can perform the same operation on multiple data points simultaneously. It is actually related when a larger entity is subdivided into various parts. For example, a plot is broken into yards for much finer granularity than just saying a plot. Attention reader! Writing code in comment? Please use ide.

Our Approach to Testing a Large-Scale C++ Codebase

Feature-based opinion mining from product reviews is a difficult task, both due to the high semantic variability of opinion expression, as well as because of the diversity of characteristics and sub-characteristics describing the products and the multitude of opinion words used to depict them. Further on, this task supposes not only the discovery of directly expressed opinions, but also the extraction of phrases that indirectly or implicitly value objects and their characteristics, by means of emotions or attitudes. Last, but not least, evaluation of results is difficult, because there is no standard corpus available that is annotated at such a fine-grained level and no annotation scheme defined for this purpose.

Skip to Main Content. A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. Use of this web site signifies your agreement to the terms and conditions.

Skip to Main Content. A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. Use of this web site signifies your agreement to the terms and conditions.


Granularity also called graininess , the condition of existing in granules or grains , refers to the extent to which a material or system is composed of distinguishable pieces. It can either refer to the extent to which a larger entity is subdivided, or the extent to which groups of smaller indistinguishable entities have joined together to become larger distinguishable entities. Coarse-grained materials or systems have fewer, larger discrete components than fine-grained materials or systems. The concepts granularity , coarseness , and fineness are relative; and are used when comparing systems or descriptions of systems.

О его существовании знали только три процента американцев. - АНБ, - пошутил приятель, - означает Агентство, которого Никогда не Было. Со смешанным чувством тревоги и любопытства Беккер принял приглашение загадочного агентства. Он проехал тридцать семь миль до их штаб-квартиры, раскинувшейся на участке площадью тридцать шесть акров среди лесистых холмов Форт-Мида в штате Мэриленд. После бесчисленных проверок на контрольно-пропускных пунктах он получил шестичасовой гостевой пропуск с голографическим текстом и был препровожден в роскошное помещение, где ему, как было сказано, предстояло вслепую оказать помощь Отделению криптографии - элитарной группе талантливых математиков, именуемых дешифровщиками.

Потребление энергии на среднем уровне. Последний шифр, введенный в ТРАНСТЕКСТ… - Она замолчала. - Что .

Беккер достал блокнот. - Итак, начнем с утра. Расскажите мне, что произошло.

Нуматака тоже был уверен, что компания это сделает. В эпоху цифровой связи понятие неприкосновенности частной жизни ушло в прошлое. Записывается. Телефонные компании могут сообщить, кто вам звонил и как долго вы говорили. - Сделайте это, - приказал .

 - Ни один из новых шифрованных файлов нельзя вскрыть без ТРАНСТЕКСТА. Вероятно, Цифровая крепость - это стандартный алгоритм для общего пользования, тем не менее эти компании не смогут его вскрыть. - Это блистательная рекламная операция, - сказал Стратмор.

 Боюсь, что. И мы должны его найти.