Combining Generality and Practicality in a Conit-Based Continuous ConsistencyModel for Wide
Combining Generality and Practicality in a Conit-Based Continuous Consistency
Model for Wide-Area Replication
Haifeng Yu Computer Science Department Duke University Durham,NC27708-0129
yhf@https://www.wendangku.net/doc/355997838.html,
Amin Vahdat Computer Science Department Duke University Durham,NC27708-0129
vahdat@https://www.wendangku.net/doc/355997838.html,
Abstract
Replication is a key approach to scaling wide-area ap-plications.However,the overhead associated with large-scale replication quickly becomes prohibitive across wide-area networks.One effective approach to addressing this limitation is to allow applications to dynamically trade re-duced consistency for increased performance and availabil-ity.Although extensive study has been performed on relaxed consistency models in traditional replicated databases, none of the models can simultaneously achieve the fol-lowing two typically con?icting requirements imposed by wide-area applications:generality(capturing application-speci?c consistency semantics)and practicality(enabling ef?cient application-independent consistency protocols to be designed and providing natural ways to express appli-cation semantics).
In this paper,we propose a conit-based continuous con-sistency model designed to simultaneously achieve general-ity and practicality.Our conit theory provides generality, where application-speci?c consistency requirements are ex-ported as conits.Practicality is achieved by using a simple, spanning set of metrics for conit consistency and by using a per-write weight speci?cation.We demonstrate the gen-erality of our model through representative wide-area ap-plications and by showing that a number of existing mod-els can be expressed as instances of our model.Our ef-?cient,application-independent consistency protocols and prototype implementation verify its practicality.
1.Introduction
Replication is a key approach to scaling wide-area ap-plications,such as e-commerce systems,dynamic content distribution,wide-area collaborative applications,sensor networks,and electronic bulletin boards.At the same time,the overhead associated with strong consistency in large-scale replication quickly becomes prohibitive across wide-area networks.One effective approach to addressing this limitations to allow these applications to dynamically trade reduced consistency for increased performance and availability[5,10,14,24,35,34]based on a continuous(re-laxed)consistency model.This diverse class of wide-area applications imposes the following two typically con?icting requirements,generality and practicality,on the continuous consistency model:
Generality These applications have rich and application-speci?c consistency semantics.For example,a shared editor may have well-de?ned,but totally different con-sistency semantics from an inventory maintenance system for e-commerce.Thus the consistency model must be suf-?ciently general and abstract to capture a wide range of consistency semantics.
Practicality The wide applicability of Internet data replica-tion requires the model to be practical to use in regular application design.More speci?cally,by practicality,we mean i)in spite of application-speci?c semantics,the pro-tocols maintaining such consistency semantics should be application-independent and highly-ef?cient and ii)the way that consistency semantics are expressed must be natural and easy to use.
The goals of generality and practicality typically con?ict with one another.One effective approach to achieve gener-ality is to avoid de?ning a uniform consistency model for all applications.Instead,applications are allowed to spec-ify their own consistency semantics.However,the consis-tency protocols enforcing such a model typically cannot be optimized in an application-independent manner.Also,to capture arbitrary semantics,the model has to be abstract, providing no natural ways for application programmers to use the model in many cases.
Figure1.How the conit-based consistency
model achieves two typically con?icting
goals.
In the context of traditional replicated databases,much research[2,3,4,8,9,10,18,19,20,24,25,26,30,31,32] has been performed on relaxed consistency models.How-ever,such traditional models typically achieve only one of generality and practicality.Some of the consistency mod-els[2,19,20,31]are general enough to allow a wide range of applications to express their consistency semantics. However,they provide no practical,ef?cient,application-independent protocols to enforce the model and no natural API for application programmers,thus failing to meet the practicality requirement.Other relaxed consistency models [3,4,8,9,10,18,24,25,26,30,32]have easy to use inter-faces and can be ef?ciently implemented,but they typically address the consistency requirements of a speci?c class of applications.
In this paper,we propose a conit-based continuous con-sistency model for wide-area data replication to simultane-ously achieve generality and practicality(Figure1).Gener-ality is achieved by our conit theory.Each conit logically represents one particular application-speci?c consistency requirement.For example,in a replicated bulletin board, one possible consistency requirement is to bound the num-ber of messages posted by other users but not seen locally. Another requirement can be the ordering among displayed news messages.These requirements serve as the de?nitions of conits.Consistency is de?ned on conits rather than data items and each conit has an application-independent con-sistency level.Each access(i.e.read or write)speci?es the required consistency level for each conit it depends on.
Practicality is achieved in our model by i)using a simple, spanning set of metrics for conit consistency and ii)express-ing semantics by simply specifying per-write weights.The ?exibility of conits allows application-speci?c consistency semantics to be“absorbed”by the conit de?nition layer. Thus we can use a simple,application-independent,span-ning set of metrics to de?ne conit consistency,which en-ables the design of highly-ef?cient application-independent consistency maintenance protocols.To provide natural API to application programmers,we avoid the complexity of exporting abstract conit de?nitions.Instead,each access speci?es the conit set it depends on and each write carries information about how it affects the consistency of each conit.The application programmer thus uses the model by attaching the necessary information to each access and never needs to explicitly de?ne conits.Our prototype imple-mentation and three sample application(replicated bulletin board,airline reservation and QoS load balancing for web servers)[34]have veri?ed the practicality and scalability of the model.
The rest of this paper is organized as follows.Section 2describes the replication model we assume.In Section3, we present the conit-based continuous consistency model and discuss its generality and practicality.Next,Section4 further explores the generality of our model through some representative wide-area applications and studies how some previous relaxed consistency models can be expressed as special instances of our model.An overview of the proto-cols implementing our consistency model is given in Sec-tion5.Finally,Section6places our work in the context of related work and Section7presents our conclusions.
2.System model
Application data,referred to as the database for simplic-ity,is replicated in full at multiple sites.Each replica ac-cepts logical reads and writes from users that may consist of multiple primitive read/write operations.Writes in our model are procedures and replicas maintain consistency by propagating write procedures(rather than the data written) as in Bayou[23]and N-ignorant systems[18].A write pro-cedure checks for con?icts with the underlying database be-fore updating the database state,allowing for application-speci?c con?ict checking in a relaxed consistency environ-ment.In case of a con?ict,a write procedure may take an alternative action.
Each replica maintains a write log,containing all writes applied to its database image.Furthermore,each replica uses standard concurrency control mechanisms to ensure lo-cal serializability.The replica that?rst accepts an access (i.e.,read or write)from a client is called the originating replica for that access.All other replicas are remote repli-cas.When?rst applied to a replica,a write is in a tenta-tive state and returns an observed result to the user.The write can then be propagated to other replicas.Writes in a replica’s write log may be reordered,e.g.rolled-back and then re-applied in a different order,with potentially differ-ent results.Write reordering is assumed to be isolated from reads and writes.At some point,a write becomes commit-
Physical World Logical World Figure2.Role of the conit theory.
ted,which means it will never be reordered again.The ac-tual result of a write is thus de?ned to be its return value when?nally committed.Reads are processed once and are never reordered.The observed result of a read is the value returned to a client query,while its actual result is the value that should be returned to the user if1SR with external order (de?ned below)were maintained.
The traditional de?nition of strong consistency for repli-cated data is one-copy serializability(1SR)[6].However, the lack of timing information in1SR makes it inappropriate for Internet applications.For example,in replicated stock quotes systems,stale values are allowed to be read even if 1SR is maintained,those reads can be considered to execute “in the past”by1SR.As in timed consistency[28,29]and external consistency[1],we augment1SR with external or-der,which is a partial order over all accesses.An access externally precedes another access if returns its ob-served result to the user(in strict wall-clock time)before
is submitted to its originating replica.We say an execution on replicated data is1SR with external order(1SR+EXT) if the execution is equivalent to a serial execution that is compatible to external order.Hereafter,we equate“strong consistency”with1SR+EXT.
3.Conit-based continuous consistency model
In this section,we?rst present the conit theory and ex-plain how it supports the consistency semantics of a broad range of replicated wide-area applications.Next,we for-mally de?ne conits and their application-independent con-sistency level.We?nish this section by demonstrating how applications can specify their consistency semantics by sim-ply assigning weights to individual write operations.3.1.Conit theory,application semantics and conit
consistency
Applications observe consistency from the results of reads and writes.With strong consistency,the observed result always equals its actual result.As we relax consis-tency,the observed result and the actual result begin to di-verge.The meaning of the difference to the end users de-pends on application semantics.Thus,in order to quantify consistency and capture the semantic discrepancy between observed and actual results,we believe that a pre-de?ned uniform consistency model is inappropriate.Instead,the consistency model should allow the application to export its speci?c consistency requirements,so that the model can ad-dress the consistency semantics that the application is sen-sitive to.
The approach we adopt in our model is to allow appli-cations to de?ne each consistency requirement as a conit. For example,in a replicated bulletin board,sample consis-tency requirements include:i)the difference between ob-served/actual number of messages,ii)the number of out-of-order messages in the current view,and iii)the consis-tency of messages posted by friends.These requirements can all serve as conit de?https://www.wendangku.net/doc/355997838.html,ing these conit de?ni-tions,our conit theory maps the physical world,composed of the physical database together with the reads and writes operating on physical data items,to a logical world(Figure 2).The logical world contains a semantics-base,consist-ing of application-speci?c consistency semantics(conits), and reads/writes conceptually operating on the semantics. Here a read/write depends on the conits with which it is concerned,and conits are affected by writes.The semantic difference between the observed and actual return value of an access is then solely determined by the depend-on conit set.For example,suppose we de?ne a conit to capture the consistency of messages posted by a user’s friends.Then if the user only cares about messages posted by her friends,
the semantic difference between the observed and actual re-sult of a read is solely determined by that conit.A write (message post)by a friend will affect the conit,while a write from other users has no effect on the conit.
In dealing with consistency,only the semantics-base is interesting to the application.Thus in our model,consis-tency is never speci?ed on data items,rather,each conit has a consistency level.Each access then speci?es the required consistency level for each conit it depends upon.Because the de?nition of each conit can be very?exible,we expect that the mapping between the physical world and the log-ical world can“absorb”most application-speci?c consis-tency semantics.This allows us to use a simple,application-independent,spanning set of metrics for conit consistency. We choose three metrics,Numerical Error,Order Error and Staleness,for conit consistency.Each conit has a logical nu-merical value.For example,in a bulletin board,the value of a conit could be the number of messages.Numerical error is the difference between the observed value of a conit and its actual value if strong consistency were enforced.With the previous conit de?nition,numerical error will be re?ected back to the physical world as the difference between the observed and actual number of messages.Order error is the weighted out-of-order writes(subject to reordering and changing behavior)that affect a conit.In the bulletin board example,order error is the number of out of order messages. Staleness is the age of the oldest write(globally across the system)affecting the conit that has not been seen by the local replica.Depending on conit de?nitions,these three metrics for conit consistency will translate to different ap-plication semantics.Section4will further discuss the gen-erality provided by user-de?ned conits and the meaning of these metrics in various situations.
3.2.Formal conit/consistency de?nition
We now formalize the previous discussion on conit and consistency,starting from the concept of history.A history is a totally ordered(serial)set of reads and writes.Because standard concurrency control mechanisms on each replica ensure local serializability,we can de?ne the local history of a replica to be the history corresponding to the equivalent serial execution of all accesses processed by that replica. The local histories are subject to reordering(due to write reordering).Causal order is a partial order de?ned over all accesses.An access causally precedes another access if is in the local history of’s originating replica when is accepted.To de?ne a consistency spectrum,we need to use a global history that corresponds to a strongly consistent execution for reference purpose.Thus,we de-?ne ECG history(external-order-compatible,causal-order-compatible,global history)to be a history that is compati-ble with external and causal order and contains all accesses accepted by the system.Unless otherwise speci?ed,the fol-lowing discussion de?nes the consistency spectrum as the
distance between local histories and a particular ECG his-tory.
We use to denote the database state at a particular time.De?ne to be the initial state of the database.
The notation denotes the database state after apply-ing write procedure to database state,while
means the database state after applying all writes in history (in history order)to.For each access,its observed pre?x history()is its originating replica’s local
history when the access is submitted.
is called the access’s observed database state(), which determines the observed result of an access.The ac-tual pre?x history()of an access is the longest pre?x of the ECG history that does not contain that access.
is called the access’s actual database state(),which determines the actual result of the ac-cess.The difference between the observed and actual result of an access is then determined by the“difference”between and.
A conit is a function that maps a database state
to a real number.An application de?nes a conit set
,which can be in?nite,to export its con-sistency semantics.De?ne the function(numeri-cal weight)of a write,conit and database state to be.De?ne the function(order weight)to be a mapping from the tuple to a non-negative real value.To simplify discussion,we assume that and are inde-pendent of(although our model is more general),so we can use the notations and.
A write affects a conit if either or
.For a history,de?ne the write or-der projection of on a conit set(denoted by)to be the sequence of writes obtained by deleting all writes in,such that
.De?ne to be the longest common pre?x of and.
For an access depending on a conit set
consistency is de?ned for each ()and is a three-dimensional vector(Numerical Error,Order Error,Staleness)as in Figure3.In the ?gure,function()is the wall-clock time that access is submitted by(returns to)the user.
Figure4illustrates the de?nition of our three consis-
tency metrics.For simplicity,we assume that the writes do not depend upon any conit and carry unit numerical weight and unit order weight for each affected conit.In this example,the read depends on two conits,and .Since,and affect and each write has a numerical weight of one,we have
in the ECG history.On the other hand,in
Figure 3.Conit consistency metrics.
R1 dep?on W1 affect W2 affect W3 affect W4 affect W5 affect R2 dep?on
W1 affect W3 affect W4 affect W2 affect R2 dep?on ECG History
Local History on Replica1
oweight(W, F) = 1
affected by write W:nweight(W, F) = 1{F1, F2}{F3}{F1}{F3}{F2}{F1}{F1, F2}
{F1, F2}{F3}{F2}{F1}{F1, F2}
Consistency of F2 for R2:
NE(absolute) = 0 OE = 1 ST = 0
ST = stime(R2) ? rtime(W5) NE(absolute) = 1 OE = 1Consistency of F1 for R2:
For each conit F
Figure 4.Conit consistency example.
the local history of
,we have .Thus,the absolute numerical error
of is and staleness is .For order error,from the ECG history,we know that
.In the local his-tory,for read ,.
Thus,the order error for
is .Similarly,the consistency
of for read is .
To choose a consistency level,the application speci?es bounds for the three metrics on a per-access and per-conit basis.Consistency is properly maintained if an ECG his-tory,,exists such that the numerical error,order error and staleness of each (access,conit)tuple are within bounds with respect to .The following theorem ensures that the result of each access is independent of the consistency level of other accesses.This self-determination property allows the application to provide differentiated consistency qual-ity of service on a per access basis.Due to space limita-tions,the proof of this theorem and of all other theorems and corollaries are made available separately[33].
Theorem 1(Self-Determination)The semantic difference between the observed result and actual result of an access is guaranteed by the consistency level of the access,inde-pendent of the consistency of other accesses.
Proof :Directly from the de?nition of the consistency of an access.
3.3.Extremes of the continuous consistency model
Tuning bounds on numerical error,order error and stal-eness of each access/conit can provide different levels of consistency.To determine the range covered by our con-tinuous consistency model,we study the two extremes of
the spectrum:when the metrics are set to
and .The weak consistency extreme is achieved when
none of the metrics are bounded and the system does not impose any restrictions on execution.We will explore the properties of the strong consistency extreme of our model by studying its relationship with 1SR[6].An execution on replicated data is 1SR if it is view equivalent[6]to a serial execution on non-replicated data.An access reads from a write if reads some data item that was last written by ,and two executions are view equivalent if each access reads from the same write in the two executions.
Theorem 2The conit-based continuous consistency model produces 1SR+EXT history if the application speci?es the following consistency:
1.A conit is de?ned for each data item in the database,where is the total number of writes applied to that data item.
2.A write affects the conit set corresponding to the data items the write updates,with unit order weight.
3.Each access depends upon the conit set corresponding to the data items it reads.
4.Zero numerical error (implying zero staleness)and zero order error are enforced in all cases.Proof :By de?nition of the continuous consistency model,an ECG history exists such that numerical error and order error are zero with respect to it.This ECG history is se-rial and compatible with external order and contains all ac-cesses processed by the system.Thus,to prove the pro-duced local histories are 1SR+EXT,we only need to show they are view equivalent to the ECG history.For each data item an access reads,consider the set of writes ()that updates that data item.Since numerical error is zero
// require (3, 0, 60) on "MsgFronFriends"
DependonConit("MsgFromFriends", 3, 0, 60); AffectConit("AllMsg", 1, 1);
// unit nweight and unit oweight PostMessage(String msg) {
// this write does not depend on any conit if (I am a friend of Alice) // unit nweight and unit oweight
AffectConit("MsgFromFriends", 1, 1);ReadMessages() {
// require (10, 5, 99999) on "AllMsg" DependonConit("AllMsg", 10, 5, 9999);
Retrieve news messages;}
}
(b)
(a)
Append msg to the bulletin board;Figure https://www.wendangku.net/doc/355997838.html,ing weight speci?cation in replicated bulletin board.
and each write carries a unit numerical weight for each data item it writes,we know that the same number of writes from precedes in as in ,where
/is the observed/actual pre?x history
of .Because ECG history is compatible with causal or-der,if
()precedes in ,then precedes in .So we know that the same set of writes from precedes in as in .Since each write also carries unit order weight for each data item it writes,zero order error ensures that these writes are
in the same order in
and .So access reads from a write in iff it reads from in .Thus,the local histories are 1SR+EXT.In the last section,we discussed the self-determination of each access.Now we highlight the implications of this result for strongly consistent accesses.
Corollary 1(Self-Determination of Strongly Consistent Accesses)If a conit is de?ned for each data item and each write carries a unit numerical/order weight for each af-fected conit,then for an access requiring zero numerical error and zero order error on all conits it depends upon,the observed result equals the actual result.
Proof :Directly from Theorem 1and Theorem 2.
If we only require 1SR for the strong consistency extreme,then reads are allowed to observe non-zero numerical error:
Theorem 3The conit-based continuous consistency model produces 1SR history if the application speci?es the follow-ing consistency:
1.A conit is de?ned for each data item in the database,where is the total number of writes applied to that data item.
2.A write affects the conit set corresponding to the data items the write updates,with unit order weight.
3.Each access depends upon the conit set corresponding to the data items it reads.
4.Zero numerical error and zero order error are enforced for all conits a write depends upon.
5.Zero order error is enforced for all conits a read de-pends upon.
Proof :Again,we consider the ECG history of the exe-cution.However,because of non-zero numerical errors for reads,the produced local histories may have different read-from relations from those in the ECG history.We will construct another global history by reordering the reads in the ECG history in the following manner.For a
read depending upon conit set
,suppose is the last write in ,where
is the observed pre?x history of .Since the
ECG history is compatible with causal order,we know must precede in the ECG history.To obtain ,ev-ery in the ECG history is moved from its original place forward to the place immediately after the corresponding
last write
in .Next,we will show that the produced local histories are view equivalent to the global history .From the proof of Theorem 1,we know that a write reads from the same write in the lo-cal history as in .All writes in and are in the same order,thus a write reads from the same write in the
local history as in
.For a read ,de?ne to be the longest pre?x of that does not contain .Be-cause the order error of each conit that depends upon is zero and each write carries a unit order weight for at lease one conit,we know that
.So reads from a write
in the local history iff it reads from in .Thus,the local histories are 1SR.
3.4.Exporting conit de?nitions through weight
speci?cation
To achieve practicality,we use weight speci?cation to provide natural API for application programmers and avoid the complexity of exporting abstract conit de?nition func-tions.Recall from Section 3.2that we de?ne a conit as a function mapping database states to real numbers.How-ever,to use our model,application programmers do not need to formally,or even conceptually,de?ne such func-tions.Rather,the application programmers can follow the following conceptual steps to use our model:
1.Crystallize high-level application consistency semantics.
2.Study how each write affects such semantics and deter-mining the corresponding numerical/order weight.
https://www.wendangku.net/doc/355997838.html,e AffectConit()statements to attach numeri-cal/order weights to writes.
4.Determine the depend-on conit set and consistency level of each access according to application requirements.
5.Add DependonConit()statements to accesses to ex-press such requirements.
In the weight speci?cation step,the application directly tells the system how each write affects the return value of a conit,and the system can then infer the return value of by summing all numerical weights accumulated.The application programmers may not even be aware of the conit functions they de?ne in such a process.
Following is a concrete example of how this can be done in a replicated bulletin board.We?rst de?ne a conit with symbolic name“AllMsg”,whose value is the number of news messages,to export the consistency requirements on all news messages.Besides these semantics,a user Al-ice also de?nes another conit with a symbolic name“Ms-gFromFriends”,whose value is the number of news mes-sages posted by Alice’s friends.Thus each write has a nu-merical weight of one for each affected conit.For simplic-ity,we also use unit order weight.Figure5(a)is the message posting routine.In this example,a write does not depend on any conits and each message posted affects the conit “AllMsg”with unit numerical weight and unit order weight. If the author of the message is a friend of Alice,the message also affects the conit“MsgFromFriends”.When Alice uses the routine in Figure5(b)to retrieve news messages,she speci?es the required consistency levels for the two conits the read depends on.For example,she requires the nu-merical error,order error and staleness on conit“MsgFrom-Friends”to be within3,0and60(seconds),respectively.In this way,the actual de?nitions of the two conits“AllMsg”and“MsgFromFriends”are never directly exported to the system.Weight speci?cation can even express subjective conit de?nition functions.For example,subjective numeri-cal weight can be attached to each news message to export its relative importance.
4.Generality of the conit-based consistency
model
4.1.Exporting application semantics through conits
In this section,we argue for the utility of our approach by discussing how a number of wide-area applications can specify their consistency semantics using conits.We will notice that not all applications below can fully utilize?ne-grained continuous consistency in our model.For exam-ple,a distributed sensor system monitoring traf?c condi-tions may be interested in all possible values of staleness bounds,while a banking system may be interested in only four different staleness bounds:zero,one hour,one day and one week.Such“non-continuity”on the consistency spec-trum is inherent in the application’s semantics and a con-tinuous consistency model can only quantify consistency to the extent allowed by the applications’semantics.Also note that because our consistency model is designed to capture a wide range of semantics,not all applications below will use all three consistency metrics.
Dynamic Content Distribution Modern web services pro-duce much of their content dynamically based on database state.Consistency is a key hurdle to replicating dynamic services across the wide area.Conits address this problem by applying application-speci?c semantics to allow services to relax from strong consistency un-der certain circumstances.Consider a dynamic web page tracking the score of a football game.The application can de?ne a conit for this page and attach subjective numer-ical weights to changes in the score.For example,score changes near the end of a close game may be considered more important.Conits may further be used to limit dis-crepancies in inventory for e-commerce services or the error in stock quotes provided by?nancial services. Shared Editor We use this application to represent wide-area collaborative applications[7]In a shared editor,mul-tiple authors work on the same document simultaneously. Consistency requirements include the“amount”of modi-?cations from remote authors not seen by a user and the “instability”of the current version due to uncommitted modi?cations.Several de?nitions of conits are possible. One approach is to de?ne two conits per paragraph repre-senting the number of characters in the paragraph.One conit tracks character additions,while the other tracks deletions.Numerical error then captures the“amount”of modi?cations not seen by a user.We can de?ne the order weight of a modi?cation also to be the number of characters it affects,and order error will capture the“in-stability”of the observed version.More functionality can be provided by,for example,de?ning a conit for each (paragraph,author)pair,so that modi?cations from dif-ferent authors can have different consistency levels.Fi-nally,staleness can be used to enforce a bound on modi-?cation propagation delay.
W AN Resource Accounting/Sensor Networks These two very different applications represent a broader class of services that maintain pure numerical records that are read/updated from multiple locations.In resource ac-counting,the data records are the resource consumption of principles,while in sensor networks,the data records are the data measured by the sensors.A conit can be de-?ned for each data record or group of records with nu-merical error capturing the accuracy of the record values.
Airline Reservation System One important aspect of sys-tem consistency for this application is the percentage of reservations aborted as a result of con?icts.This aspect can be captured using numerical error in the following manner.A conit is used for each?ight and the value of the conit is de?ned to be the number of available seats on that?ight.Assuming single seat reservations(though our model is more general)and that reservations are ran-domly distributed among all available seats,the proba-bility that a reservation con?icts with another remote (unseen)reservation is. Since relative numerical error of the conit equals
,we can use to express the con?ict rate:.Thus,the sys-tem can limit the rate of reservation con?icts by bounding relative numerical error.The above formula has been ver-i?ed through experiments[34].Non-random reservation behavior will result in a higher con?ict rate,but the appli-cation may still limit con?ict rates by de?ning multiple conits over,for example,?rst class and coach seats. Distributed Games/Virtual Reality/Teleimmersion [11,14]Most of the consistency issue for these ap-plications concerns the positions and orientations of objects in the virtual world.Since both position and orientation are pure numerical data,the semantics can be easily captured by numerical error.Furthermore,using different consistency levels for each conit/access can allow differentiated focus and nimbus[5]to represent the degree of interest objects have in each other.
Traf?c Monitoring and Road Reservation Advances
in mobile technology have made“road reservation”possible.Here a mobile device is equipped to each vehicle and base stations help to collect/distributed traf?c information to allow drivers to choose the“best”route. Road reservation helps to avoid the situation where many drivers choose the same“best”route and suddenly the route becomes over-crowded.Consistency here is the accuracy of the traf?c/reservation information.We can de?ne each section of the road to be a conit,its value being the number of vehicles in that section.To be more precise,different weights can be assigned to different vehicles to take into account the vehicle size,etc. Abstract Data Types Abstract data types naturally?t into our consistency model.For example,consider a set(or hashtable)with methods add(),remove(),size() and contains().We can de?ne a conit whose value is the number of elements in the set.The accuracy of the return value of size()can then be re?ected in the numerical error of the conit.Similarly,the probability of contain()returning a correct value is determined by the numerical error.4.2.Relationship to other consistency models
To further demonstrate the generality of our conit-based consistency model,in the following,we will discuss how some previous relaxed consistency models can be expressed as special instances of our model.
Con?ict Matrix[4,8,30]The use of a con?ict matrix is a well-studied technique for relaxing the consistency of ab-stract data types.Each entry in the con?ict matrix deter-mines whether two methods on the same object can pro-ceed in parallel.Our consistency model can achieve the same functionality using the following conit de?nition. Each method is considered a write.The th row of the con?ict matrix(associated with method)is assigned a conit,.For a method corresponding to the th column of the con?ict matrix,affects
iff the matrix entry is a“con?ict”entry.For each conit affected,carries a unit numerical weight.Each method depends on conit and requires zero numer-ical error.In this way,all pairs of non-con?icting method invocations can be processed in parallel,while con?icting invocations have to be processed in a manner equivalent to1SR.A correctness proof is omitted for brevity.Note that if we enforce?nite,instead of zero/in?nity,numeri-cal error for a matrix entry,we can provide the semantics of“bounded con?ict”that cannot be obtained from a con-?ict matrix.For example,a getBalance()method on a bank account is allowed to miss no more than$50de-posited by deposit()operations.
Three-level Consistency in Lazy Replication[21]Ladin et.al.propose three different consistency levels in lazy replication.A causal transaction is causally ordered to all other causal transactions,a forced transaction is totally ordered across all replicas with respect to all other forced transactions,and immediate transactions are totally or-dered across all replicas with respect to all transactions. These consistency levels can be expressed using the fol-lowing con?ict matrix regarding the three types of trans-actions,and thus can be easily captured by our model. Sample conit speci?cations are included in the following table.
Transaction Forced
(affect)(affect,
and)
Causal No con?ict
No con?ict Con?ict (dep-on)
Immediate Con?ict
Cluster Consistency[24]Cluster consistency is a two-level consistency model proposed for mobile environ-ments.In this model,data copies are partitioned into clus-ters,where consistency constraints within a cluster must be preserved while inter-cluster consistency may be vio-lated.Two kinds of operations are allowed:strict opera-tions and weak operations.The consistency requirements of these operations can again be expressed as a con?ict matrix,and thus can be captured by our model.To en-force“m-consistency”[24]for some entries in the matrix, we can allow non-zero numerical/order error for the conit corresponding to that row.
N-ignorant System[18]In an N-ignorant system,a trans-action can run in parallel with at most other transac-tions.To emulate the behavior of an N-ignorant system, we de?ne a conit whose value is the number of transac-tions applied to the database.A system bounding numer-ical error within will behave the same as an N-ignorant system.
Timed Consistency/Delta Consistency[28,29]These models address the lack of timing in traditional con-sistency models such as sequential consistency.They require the effect of a write to be observed everywhere within time.These timed models can be readily expressed using the staleness metric on conits.
Quasi-copy Caching and its Generalization[3,10] Quasi-copy caching proposes four coherency conditions: delay condition,frequency condition,arithmetic con-dition and version condition.Delay condition imposes an upper bound on propagation delay for a data item, which is a special case of staleness on conits.Frequency condition requires the copies of a data item to be syn-chronized every seconds.We believe in most cases, frequency condition can be more ef?ciently achieved by bounding staleness.Arithmetic condition bounds the difference between copies of numerical data items, which can be captured by the numerical error on conits. The last condition,version condition,bounds the version difference among copies.It can be achieved by using a conit whose value is the number of updates applied to a data item and by bounding the absolute numerical error of the conit.Quasi-copy caching is later generalized and a few more coherency conditions proposed[10]. Due to space limitations,we will only discuss the more distinct one,object condition.This condition requires the copies of an object to be synchronized when:i)at least sub-objects of have been modi?ed,ii)at least percent of the sub-objects of have been modi?ed,or iii)sub-object of has been modi?ed.Emulating this condition requires three conits(,and),one for each of the three cases,to be de?ned for each object.
The value of both and is the number of modi?ed sub-objects.To enforce case i)and case ii),we bound the absolute error of within and bound the relative error of within.The value of the third conit is the number of updates on sub-object.The numerical error of is bounded to zero,which means updates on will incur synchronization right away.
Memory Consistency Models in Multi-Processors [16,22,27,36]Numerous number of memory con-sistency models have been proposed in the context of multi-processor/distributed shared memory.Due to space limitations,here we cannot discuss those models individually and can only give a high-level abstract discussion.Most of the consistency models are de?ned by imposing ordering requirements on load,store and other synchronization(e.g.fence,barrier,lock acquire) instructions.The system is allowed to re-order instruc-tions,that is,violate program order,during execution, as long as the ordering imposed by the consistency model is preserved.Although our consistency model cannot be directly used in computer architecture,the conit concept can still be applied.Consider a consistency model for multi-processor and a program running under this model.The ordering requirements imposed by the model on the program can always be viewed as a DAG, whose nodes and edges are instructions and ordering among instructions,respectively.To apply the conit theory,we need to rede?ne the reference history to be the history compatible to the DAG.Next,we assign a conit for each edge in the DAG.Each node in the DAG is modeled as a write and it depends on/affects the conit set corresponding to the set of incoming/outgoing edges in the https://www.wendangku.net/doc/355997838.html,st,we enforce zero numerical error on all conits.It can then be shown that the resulting model is equivalent to the original memory consistency model.A correctness proof is omitted for brevity.
5.Implementation of the continuous consis-
tency model and scalability issues
We have designed application-independent protocols to enforce conit consistency.Because of the simplicity of conit consistency metrics,the protocols can be highly op-timized.We only give an overview here,detailed dis-cussion of the protocols and their implementation can be found in[34,35].The absolute/relative numerical error bounding algorithms[35]for pure numerical data items are adopted for bounding numerical error of conits.Order error can be bounded with a write commitment algorithm,that is,an algorithm that allows replicas to agree on a write order[12,13,15,23].Staleness bounds can be enforced through a straightforward write pulling algorithm.
All our protocols for bounding the three metrics are scal-able relative to the number of conits.Such scalability is crucial for our model because the number of conits can be very large(on the order of the number of data items in the database),depending on application semantics.In numerical error bounding protocols,we avoid maintaining constant-size bookkeeping information for each conit.In-stead,such information is dynamically created when neces-sary and deleted when no longer in use.In the write com-mitment algorithms,scalability can be achieved by ignoring order relaxations enabled by multiple conits.In the extreme, if we simply use a conventional write commitment algo-rithm to generate a total order on all writes,the overhead incurred will be independent of the total number of conits. Our staleness bounding algorithm,by nature,is insensitive to the number of conits.
Finally,a system prototype and wide-area evaluation of three sample applications(replicated bulletin board,airline reservation and QoS load distribution for web servers)[34] demonstrates the practicality of our approach.
6.Related work
In[35],we propose algorithms to enforce numerical er-ror for pure numerical data records,even though the al-gorithms are also applicable to enforcing numerical error bounds for conit consistency.The prototype implementa-tion and performance data of our consistency model are pre-sented in[34].However,in[34]we focus on how consis-tency can be traded for performance and no formal de?ni-tion of conit,conit consistency or conit theory is provided. This paper concentrates on formal aspects of our consis-tency model and discusses how generality and practicality can be simultaneously achieved.
Most of the previous relaxed consistency models were not designed for the dual goals of generality and prac-ticality.Agrawal et.al.[2]propose semantics-based con-sistency criteria using guarded actions,which are primi-tive reads/writes associated with arbitrary consistency as-sertions.Wong et.al.[31]apply similar ideas to abstract data types.In their model,a history is consistent if the assertions are satis?ed when the system executes the asso-ciated read/write.In the similarity model[19,20],appli-cations de?ne certain database states to be indistinguish-able for concurrency control purposes.These three mod-els can capture a broad range of application semantics. However,they place a signi?cant burden on the applica-tion to match the model to their requirements.Further, they do not provide any practical,ef?cient protocols to en-force the requested consistency level in the general case. On the other hand,quasi-copy caching[3,10],N-ignorant systems[18],delta consistency[28],timed consistency[29], cluster consistency[24]and models based on a con?ict ma-trix for abstract data types[4,8,30]have developed ef?-cient application-independent protocols to enforce the re-laxed consistency model.However,because they use a uni-form consistency model for all applications,generality is sacri?ced in favor of the consistency requirements of a spe-ci?c class of applications.In Section4.2,we showed that all these models can be expressed using our conit-based consis-tency model.
Pu et.al.[26]propose the concept of epsilon-serializability(ESR)to relax serializability and algorithms [9,25,32]have been developed to enforce ESR.Relative to ESR,our conit-based model allows a broader range of application semantics to be expressed through?exible conit de?nitions.Another fundamental difference is that while we focus on trading consistency for reduced wide-area communication among replicas,ESR aims to increase the concurrency at a single site.The lifetime-based mutual consistency detection mechanism[17]can provide several discrete mutual consistency levels for different objects. Their mechanism is targeted to a different problem from ours,that is,to determine mutual consistency of objects in a system where client caches may retrieve individual objects from servers.Because replicas directly propagate writes in our system model,mutual consistency among data items in our model is always ensured.
7.Conclusions
In this paper,we propose a conit-based continuous con-sistency model to address the inherent overheads associated with large-scale replication in the Internet.Our model si-multaneously achieve generality and practicality.These two goals usually con?ict because generality requires applica-tion semantics to be exported,which typically precludes natural API and ef?cient,application-independent consis-tency protocols.Generality in our model is achieved by using user-de?ned conits to map the physical world to a logical world.We study the generality of our model by dis-cussing how representative wide-area applications can ex-port application-speci?c consistency semantics and how a number of existing relaxed consistency models can be ex-pressed using our model.Practicality in our model is pro-vided by i)using simple conit consistency metrics to allow application-independent consistency protocols and ii)using weight speci?cation to simplify semantics expression.A number of ef?cient,application-independent protocols en-forcing the consistency model and the prototype implemen-tation verify its practicality.
References
[1]A.Adya,R.Gruber,B.Liskov,and U.Maheshwari.Ef?-
cient Optimistic Concurrency Control Using Loosely Syn-
chronized Clocks.In Proceedings of the ACM SIGMOD Conference on Management of Data,May1995.
[2]D.Agrawal, A.E.Abbadi,and A.K.Singh.Consis-
tency and Orderability:Semantics-Based Correctness Crite-ria for Databases.ACM Transactions on Database Systems, September1993.
[3]R.Alonso,D.Barbara,and H.Garcia-Molina.Data Caching
Issues in an Information Retrieval System.ACM Transac-tions on Database Systems,September1990.
[4]B.R.Badrinath and K.Ramamritham.Semantics-based
Concurrency Control:Beyond Commutativity.ACM Trans-actions on Database Systems,March1992.
[5]S.Benford,L.Fahlen,C.Greenhalge,and J.Bowers.Man-
aging Mutual Awareness in Collaborative Virtual Environ-ments.In Proceedings of the ACM Conference on Virtual Reality and Technology,1994.
[6]P.A.Bernstein,V.Hadzilacos,and N.Goodman.Concur-
rency Control and Recovery in Database Systems.Addison-Wesley,1987.
[7]P.Dewan,R.Choudhary,and H.Shen.An Editing-based
Characterization of the Design Space of Collaborative Appli-cations.Journal of Organizational Computing,pages219–240,1994.
[8]L.B.C.DiPippo and V.F.Wolfe.Object-based Semantic
Real-time Concurrency Control.In Proceedings of the IEEE Real-Time Systems Symposium,1993.
[9]P.Drew and C.Pu.Asychronous Consistency Restoration
under Epsilon Serializability.In Proceedings of the28th Hawaiian International Conference on Systems Sciences, January1995.
[10]R.Gallersdorfer and M.Nicola.Improving Performance in
Replicated Databases through Relaxed Coherency.In Pro-ceedings of the21st International Conference on Very Large Databases,1995.
[11]L.Gautier and C.Diot.Design and Evaluation of MiMaze,a
Multi-Player Game on the Internet.In Proceedings of IEEE Multimedia Systems Conference,June1998.
[12]R. A.Golding.A Weak-Consistency Architecture for
Distributed Information https://www.wendangku.net/doc/355997838.html,puting Systems, 5(4):379–405,Fall1992.
[13]J.Holliday,R.Steinke,D.Agrawal,and A.E.Abbadi.Epi-
demic Quorums for Managing Replicated Data.In Proceed-ings of the19th IEEE International Performance,Comput-ing,and Communications Conference,February2000. [14]Y.Honda,K.Matsuda,J.Rekimoto,and R.Lea.Virtual So-
ciety:Extending the WWW to Support a Multi-User Interac-tive Shared3D Environment.In Proceedings of the First An-nual Symposium on the Virtual Reality Modeling Language, 1995.
[15]P.Keleher.Decentralized Replicated-Object Protocols.In
Proceedings of the18th Annual ACM Symposium on Princi-ples of Distributed Computing,April1999.
[16]P.Keleher,A.L.Cox,and https://www.wendangku.net/doc/355997838.html,zy Release
Consistency for Software Distributed Shared Memory.In Proceedings of the19th Annual International Symposium on Computer Architecture,May1992.
[17]R.Kordale and M.Ahamad.A Scalable Technique for Im-
plementing Multiple Consistency Levels for Distributed Ob-jects.In Proceedings of the16th IEEE International Confer-ence on Distributed Computing Systems,1996.[18]N.Krishnakumar and A.Bernstein.Bounded Ignorance:A
Technique for Increasing Concurrency in a Replicated Sys-tem.ACM Transactions on Database Systems,19(4),De-cember1994.
[19]T.-W.Kuo and A.K.Mok.Application Semantics and Con-
currency Control of Real-Time Data-Intensive Applications.
In Proceedings of the IEEE Real-Time Systems Symposium, 1992.
[20]T.-W.Kuo and A.K.Mok.SSP:A Semantics-Based Proto-
col for Real-Time Data Access.In Proceedings of the IEEE Real-Time Systems Symposium,1993.
[21]https://www.wendangku.net/doc/355997838.html,din,B.Liskov,L.Shirira,and S.Ghemawat.Providing
Availability Using Lazy Replication.ACM Transactions on Computer Systems,10(4):360–391,1992.
[22]D.Patterson and https://www.wendangku.net/doc/355997838.html,puter Architecture:A
Quantitative Approach.Morgan Kaufman Publishers,Inc., second edition,1996.
[23]K.Petersen,M.Spreitzer,D.Terry,M.Theimer,and A.De-
mers.Flexible Update Propagation for Weakly Consistent Replication.In Proceedings of the16th ACM Symposium on Operating Systems Principles(SOSP-16),pages288–301, October1997.
[24]E.Pitoura and B.K.Bhargava.Maintaining Consistency of
Data in Mobile Distributed Environments.In Proceedings of the15th International Conference on Distributed Computing Systems,May1995.
[25]C.Pu,W.Hseush,G.E.Kaiser,K.-L.Wu,and P.S.Yu.Dis-
tributed Divergence Control for Epsilon Serializability.In Proceedings of the International Conference on Distributed Computing Systems,1993.
[26]C.Pu and A.Leff.Replication Control in Distributed Sys-
tem:An Asynchronous Approach.In Proceedings of the ACM SIGMOD Conference on Management of Data,May 1991.
[27]X.Shen,Arvind,and https://www.wendangku.net/doc/355997838.html,mit-reconcile Fences
(CRF):A New Memory Model for Architects and Compiler Writers.In Proceedings of the26th Annual International Symposium on Computer Architecture,May1999.
[28]A.Singla,U.Ramachandran,and J.Hodgins.Temporal No-
tions of Synchronization and Consistency in Beehive.In Pro-ceedings of the9th ACM Symposium on Parallel Algorithms and Architectures,June1997.
[29]F.Torres-Rojas,M.Ahamad,and M.Raynal.Timed Consis-
tency for Shared Distributed Objects.In Proceedings of the 18th ACM Symposium on Principle of Distributed Comput-ing,May1999.
[30]https://www.wendangku.net/doc/355997838.html,mutativity-Based Concurrency Control for
Abstract Data Types.IEEE Transactions on Computers,De-cember1988.
[31]M.H.Wong and D.Agrawal.Tolerating Bounded Incon-
sistency for Increasing Concurrency in Database Systems.
In Proceedings of the11th Symposium on Principles of Database Systems,June1992.
[32]K.-L.Wu,P.S.Yu,and C.Pu.Divergence Control for
Epsilon-Serializability.In Proceedings of8th International Conference on Data Engineering,February1992.
[33]H.Yu and https://www.wendangku.net/doc/355997838.html,bining Generality and Prac-
ticality in a Conit-Based Continuous Consistency Model for Wide-Area Replication.Technical Report CS-2000-09,
Duke University,Computer Science Department,Durham, NC,2000.See https://www.wendangku.net/doc/355997838.html,/?yhf/ icdcstr.pdf.
[34]H.Yu and A.Vahdat.Design and Evaluation of a Continuous
Consistency Model for Replicated Services.In Proceedings of the4th Symposium on Operating Systems Design and Im-plementation(OSDI),October2000.
[35]H.Yu and A.Vahdat.Ef?cient Numerical Error Bounding
for Replicated Network Services.In Proceedings of the26th International Conference on Very Large Databases,Septem-ber2000.
[36]M.J.Zekauskas94,W.A.Sawdon,and B.N.Bershad.Soft-
ware Write Protection for Distributed Share Memory.In Pro-ceedings of the First USENIX Symposium on Operating Sys-tem Design and Implementation,November1994.
智能存包柜(储物柜)产品技术说明书
条码式寄存柜（储物柜）产品技术说明 一、适用范围 自助行李寄存柜，用于高效率、规范化且消费群体较密集的高档公共场所，如：超市、学校、机场、俱乐部、游泳馆、健身房、超市、大卖场、商场、事业单位等。客户可根据自身特点选择不同类型的机组，有条码型、指纹型、IC卡、条形码卡、磁卡、射频卡以及TM扣等，或者使用客户各自所持有的会员卡。 二、基本组成 自助行李寄存柜，是由嵌入式计算机进行控制，具有管理功能的小件物品寄存系统。主要由输入设备（条码阅读器）、输出设备（如中文液晶显示）、嵌入式计算机处理中心、管理软件以及电源组成。 三、结构、材料、钣金说明 寄存柜柜体选用0.8mm以上优质冷轧板，经冷加工成形后，用二氧化碳气体保护焊焊接装配而成，柜体结构坚固结实。箱体表面经除锈、除油、打磨、磷化处理后喷塑，塑面的颜色可由用户选定。?箱门背面增加纵向加强筋，提高箱门的防撞击能力；能有效防止和降低使用者因疏忽碰伤、磕伤，电控锁采用360度具有防撬、带防软片插入装置。 四、元器件说明 1)电控锁：经过10万次寿命测试无故障，性能可靠稳定。部件的表面均作热 涂锌防锈处理，装配成形后每把锁均作电气绝缘试验和开启试验，并按加工批量的1％作寿命试验。 2)扫描枪：台湾欣技C1000，性能稳定，读码速度快。 3)液晶显示屏：3.5寸彩色显示屏。例如:
表示：空箱，顾客可以存入的箱子 表示：满箱，有合法密码，箱内有物 4)打印机：自主研发使用日本精工打印机芯。 5)控制主板：自主研发生产。 6)切纸刀：自主研发生产、10万次寿命测试，性能稳定，配有走纸导向。 7)控制芯片：采用国际上著名大公司的工业级产品，有ATMEL公司、INTEL公 司、日立公司、德州公司等。 8)键盘：选用电话机水晶键盘，工作可靠，永不退色，使用寿命长。 五、常规技术指标 （1）功率－待机：25W 开箱：60W （2）电源电压：AC110V—AC240V 50HZ （3）后备电源（可选）：18V 7AH （4）打印速度：35mm/s （5）打印机寿命：50Km （6）柜体冷轧钢板厚度：0.8mm （7）读码速度：<0.4s （8）显示分辨率：3.5寸彩色液晶屏（可定制显示LOGO） 六、条码寄存柜主要功能说明 ?大屏幕液晶显示屏 操作步骤提示、箱门5种状态显示、时钟显示、常规故障指示、设置菜单、管理界面、工作状态显示等。 在现场安装就位后，确认各个部件正常后，即可上电开机，完整的显示： ?全开放式中文设置菜单（可定制多国语言、语音） 寄存柜的各个管理参数全部面向客户开放，客户可选择相应的菜单进行设置。设有三级管理密码，方便不同级别的管理要求。 ?快捷管理员管理菜单 在管理模式下，管理快捷菜单可应急开启箱门、清除箱门ID、查询箱门状态等操作。可设置500位管理员进行管理,每个管理员可根据自己的用户号及管
二、分析天平的基本操作
模块二滴定分析基本操作任务一分析天平的基本操作 一、分析天平的分类 分析天平是定量分析中最常用的准确称量物质的仪器。分析天平分类：等臂（双盘）分析天平、不等臂（单盘）分析天平、电子天平。 二、电子天平的介绍 1、工作原理：电磁平衡原理，秤盘通过支架连杆支架作用于线圈上，重力方向向下。线圈内有电流通过时，根据电磁基本理论，通电的导线在磁场中将产生一个向上作用的电磁力，与秤盘重力方向相反大小相同，与之相平衡，而通过导线的电流与被称物体的质量成正比。 2、性能特点：a.使用寿命长，性能稳定，灵敏度高，体积小，操作方便 b.称 量速度快、精度高 c.具有自动校准、累计称量、超载显示、自动去皮等功能 3、称量的一般程序 水平调节——打扫——预热——开启显示器——校准——称量——结束工作 ①水平调节检查水平仪，调节水平调节脚，使水泡位于水平仪中心。 ②打扫打扫天平秤盘 ③预热通电预热30min以上 ④开启显示器按ON键，显示器亮，显示屏出现0.0000g ⑤校准按“校准”键 ⑥称量被称物置于秤盘中间进行称量 ⑦称量结束工作取下被称物，核对零点，关闭天平，进行使用登记 4、基本称量方法 ①直接称量法将称量物直接放在天平盘上直接称量物体的质量。例如，称量 小烧杯的质量，容量器皿校正中称量某容量瓶的质量，重量分析实验中称量某坩埚的质量等，都使用这种称量法。 ②固定质量称量法
用于称量某一固定质量的试剂(如基准物质)或试样。适于称量不易吸潮、在空气中能稳定存在的粉末状或小颗粒样品。 A、去皮将干燥的容器置于秤盘上，待显示平衡后按“去皮”键扣除皮重并显示零点 B、加样打开天平门，用药匙将试样抖入容器内，使之达到所需质量。 固定质量称量法注意：若不慎加入试剂超过指定质量，用牛角匙取出多余试剂，直至试剂质量符合指定要求为止。严格要求时，取出的多余试剂应弃去，不要放回原试剂瓶中。操作时不能将试剂散落于天平盘等容器以外的地方，称好的试剂必须定量地由表面皿等容器直接转入接受容器，此即所谓“定量转移”。 ③递减称量法（减量法） 用于称量一定质量范围的样品或试剂。样品易吸水、易氧化或易与二氧化碳等反应时，可选择此法。。 称量步骤：试样的保存——取出盛试样的称量瓶——称出称量瓶质量——敲样——再称出其质量——样品质量——连续称样——称量工作结束 A、试样保存待称样品放于洁净的干燥容器（称量瓶）中，置于干燥器中保存 B、取出称量瓶左手戴手套取出称量瓶或者用折叠成约1cm的纸取出 C、称出称量瓶质量称出称量瓶质量，记录数据 D、敲样将称量瓶取出，在接收容器的上方倾斜瓶身，用称量瓶盖轻敲瓶口上部使试样慢慢落入容器中，瓶盖始终不要离开接受器上方。当倾出的试样接近所需量时，一边继续用瓶盖轻敲瓶口，一边逐渐将瓶身竖直，使粘附在瓶口上的试样落回称量瓶，然后盖好瓶盖，准确称其质量。两次质量之差，即为试样的质量。按上述方法连续递减，可称量多份试样。
精神分裂症的病因及发病机理
精神分裂症的病因及发病机理 精神分裂症病因：尚未明，近百年来的研究结果也仅发现一些可能的致病因素。（一）生物学因素1．遗传遗传因素是精神分裂症最可能的一种素质因素。国内家系调查资料表明：精神分裂症患者亲属中的患病率比一般居民高6.2倍，血缘关系愈近，患病率也愈高。双生子研究表明：遗传信息几乎相同的单卵双生子的同病率远较遗传信息不完全相同 的双卵双生子为高，综合近年来11项研究资料：单卵双生子同病率（56.7%），是双卵双生子同病率（12.7%）的4.5倍，是一般人口患难与共病率的35-60倍。说明遗传因素在本病发生中具有重要作用，寄养子研究也证明遗传因素是本症发病的主要因素，而环境因素的重要性较小。以往的研究证明疾病并不按类型进行遗传，目前认为多基因遗传方式的可能性最大，也有人认为是常染色体单基因遗传或多源性遗传。Shields发现病情愈轻，病因愈复杂，愈属多源性遗传。高发家系的前瞻性研究与分子遗传的研究相结合，可能阐明一些问题。国内有报道用人类原癌基因Ha-ras-1为探针，对精神病患者基因组进行限止性片段长度多态性的分析，结果提示11号染色体上可能存在着精神分裂症与双相情感性精神病有关的DNA序列。2．性格特征：约40%患者的病前性格具有孤僻、冷淡、敏感、多疑、富于幻想等特征，即内向
型性格。3．其它：精神分裂症发病与年龄有一定关系，多发生于青壮年，约1/2患者于20～30岁发病。发病年龄与临床类型有关，偏执型发病较晚，有资料提示偏执型平均发病年龄为35岁，其它型为23岁。80年代国内12地区调查资料：女性总患病率（7.07%。）与时点患病率（5.91%。）明显高于男性（4.33%。与3.68%。）。Kretschmer在描述性格与精神分裂症关系时指出：61%患者为瘦长型和运动家型，12.8%为肥胖型，11.3%发育不良型。在躯体疾病或分娩之后发生精神分裂症是很常见的现象，可能是心理性生理性应激的非特异性影响。部分患者在脑外伤后或感染性疾病后发病；有报告在精神分裂症患者的脑脊液中发现病毒性物质；月经期内病情加重等躯体因素都可能是诱发因素，但在精神分裂症发病机理中的价值有待进一步证实。（二）心理社会因素1．环境因素①家庭中父母的性格，言行、举止和教育方式（如放纵、溺爱、过严）等都会影响子女的心身健康或导致个性偏离常态。②家庭成员间的关系及其精神交流的紊乱。③生活不安定、居住拥挤、职业不固定、人际关系不良、噪音干扰、环境污染等均对发病有一定作用。农村精神分裂症发病率明显低于城市。2．心理因素一般认为生活事件可发诱发精神分裂症。诸如失学、失恋、学习紧张、家庭纠纷、夫妻不和、意处事故等均对发病有一定影响，但这些事件的性质均无特殊性。因此，心理因素也仅属诱发因
电子天平使用说明书.
电子天平使用说明书 使用方法 ◎准备 1、将天平安放在稳定及水平的工作台上,避免振动、气流、阳光直射和剧烈的温度波动; 2、安装称盘; 3、接通电源前请确认当地交流电压是否与天平所附的电源适配器所需电压一致; 4、为获得准确的称量结果,在进行称量前天平应接通电源预热30分钟。 ◎电源 1. 天平随机附配交流电源适配器,输入220+22-33V ~ 50Hz 输出9V 300mA 2. 天平选用电池供电时可打开天平底部的电池盖按极性指示装入电池即可,建议使用9伏碱性电池,可连续工作约12小时。 当天平电池供电时,显示屏左上角电量指示框显示段数表明电池的状态(显示3段:电池充足,显示0段:电池耗尽,当电池电量将耗尽时,最后一个显示段闪烁。 ◎开机 在称盘空载情况下按<开/关>键,天平依次进入自检显示(显示屏所有字段短时点亮、型号显示和零状态显示,当天平显示零状态时即可进行称量; 当遇到相关功能键设置有误无法恢复时,按<开/关>键重新开机即可恢复初始设置状态。
◎校准 为获得准确的称量结果,必须对天平进行校准以适应当地的重力加速度。校准应在天平预热结束后进行,遇到以下情况必须使用外部校准砝码对天平进行校准。 1. 首次使用天平称量之前; 2. 天平改变安放位置后。 校准方法与步骤: 1.准备好校准用的标准砝码并确保称盘空载; 2.按<去皮>键:天平显示零状态; 3.按<校准>键:天平显示闪烁的CAL—XXX,(XXX一般为100、200或其它数字,提醒使用相对应的100g、200g或其它规格的标准砝码 4.将标准砝码放到称盘中心位置,天平显示CAL-XXX,等待几秒钟后,显示标准砝码的量值。此时移去砝码,天平显示零状态,则表示校准结束,可以进行称量。如天平不零状态,应重复进行一次校准工作。 ◎称量 天平经校准后即可进行称量,称量时必须等显示器左下角的“○”标志熄灭后才可读数,称量过程中被称物必须轻拿轻放,并确保不使天平超载,以免损坏天平的传感器。 ◎清零或去皮 清零:当天平空载时,如显示不在零状态,可按<去皮>键,使天平显示零状态。此时才可进行正常称量。
古代汉语词类活用例句列举
古代汉语词类活用例句列举 古代汉语词类活用例句列举《郑伯克段于鄢》１、例：壮公生，惊姜氏。P97 惊：用作使动，使。。。惊。２、例：无生民心。P99 生：用作使动，使。。。产生。３、例：若阙地及泉，隧而相见。P101 隧：名词动用。《公孙无知之乱》４、豕立而啼，P109 立：名词作状语，像人一样丫立。〈安之战〉５、皆主？献子。P117 主：名词动用，以。。。为主。６、君无所辱命。P119 辱：动词使动，使。。。受辱。７、从左右，皆肘之。P123 肘：名词使动，表示用胳膊推撞。８、臣辱戎士。123 辱：动词使动。９、人不难以死免其君。P123 免：用作使动，使。。。免于。１０、故中御而从齐候。P123 中：方位名词做状语。〈子产说范宣子轻敝〉１１、三周华不注。P122 周：
名词动用。１２、郑人病之。P129 病：名词用作意动。１３、象有齿而焚其身。P130 焚：动词用作使动。１４、宣子说，乃轻弊。P130 轻：形容词用作使动，使。。。轻。〈苏秦连横约纵〉１５、今先生俨然不运千里而庭教之。P182 远：形容词用作意动。１６、明言章理，兵甲愈起。P183 明、章：用作使动。 1 １７、辨言伟服。攻战不息。P183 辩、伟：都用作使动，使。。。雄辩，使。。。华美。１８、繁称文辞，天下不冶。P183 文：名词用作使动。１９、夫徒处而致利，安坐而广地。P183 广：形容词用作使动，使。。。广。２０、言语相结，天下为一。P183 言语：名词作状语。２１、今欲并天下，凌万乘，诎敌国。制海内，子元元。臣诸候。非兵不可。P183 诎：用作使动，使。。。屈服；子：名词用作使动，使。。。成为子女；臣：名词用作使动，使。。。成为臣子。２２、约纵散横，以抑强秦。
存包柜项目方案说明
存包柜项目方案说明 一、项目基本信息 （一）项目名称 存包柜项目 （二）项目建设单位 xxx有限公司 （三）法定代表人 于xx （四）公司简介 本公司奉行“客户至上，质量保障”的服务宗旨，树立“一切为客户着想” 的经营理念，以高效、优质、优惠的专业精神服务于新老客户。 公司依托集团公司整体优势、发展自身专业化咨询能力，以助力产业提高运营效率为使命，提供全方面的业务咨询服务。 公司将加强人才的引进和培养，尤其是研发及业务方面的高级人才，健全研发、管理和销售等各级人员的薪酬考核体系，完善激励制度，提高公司员工创造力，为公司的持续快速发展提供强大保障。
上一年度，xxx（集团）有限公司实现营业收入7026.99万元，同比增长25.92%（1446.42万元）。其中，主营业业务存包柜生产及销售收入为5749.34万元，占营业总收入的81.82%。 根据初步统计测算，公司实现利润总额1653.53万元，较去年同期相比增长375.52万元，增长率29.38%；实现净利润1240.15万元，较去年同期相比增长139.78万元，增长率12.70%。 （五）项目选址 xx新兴产业示范基地 （六）项目用地规模 项目总用地面积12559.61平方米（折合约18.83亩）。 （七）项目用地控制指标 该工程规划建筑系数71.28%，建筑容积率1.01，建设区域绿化覆盖率7.20%，固定资产投资强度188.57万元/亩。 项目净用地面积12559.61平方米，建筑物基底占地面积8952.49平方米，总建筑面积12685.21平方米，其中：规划建设主体工程7943.44平方米，项目规划绿化面积913.28平方米。 （八）设备选型方案 项目计划购置设备共计91台（套），设备购置费1292.40万元。 （九）节能分析 1、项目年用电量486324.58千瓦时，折合59.77吨标准煤。
分析天平使用说明书
TG-328分析天平使用说明书 一、分析天平结构及作用原理： 1.该分析天平，是属于双盘等臂式，横梁采用铜镍合金制成，上面装有玛瑙刀三把，中间为固定的支点刀，两边可调整的承重刀。 好仪器，好资料，尽在沧州建仪（https://www.wendangku.net/doc/355997838.html,）。欢迎查询。 打造中国建仪销售第一品牌，树立沧州产品全新形象 2.支点刀位于支点刀垫上，支点刀垫固定在天平立柱上端。 3.横梁停动装置为双层折翼式，在天平开启时，横梁上的承重刀必须比支点刀先接角触，为了避免刀锋损坏和保证横梁位置的再现性，开启天平求轻稳，避免冲击，摇晃。 4.横梁的左右两端悬挂承重挂钩，左承重挂钩上装有砝码承重架，该二零碎件分别挂在小刀刃上，另有秤盘各一件分别挂在承重挂钩上。 5.整个天平固定在大理石的基座板上，底板前下部装有二只可供调整水平位置的螺旋脚，后面装有一只固定脚，天平木框前面有一扇可供启闭及随意停止在上下位置的玻璃门，右侧有一扇玻璃移门。 6.秤盘上节中间的阻尼装置，是用铝合金板制成，固定在中柱上。是利用空气阻力来减少横梁的摆动时间，达到静止迅速，从而提高工作效率。
7.光学报影装置，固定在底板上前方，可直接读出0.1-10毫克以内的重量值。 8.天平外框左侧装有机械加码装置，通过三档增减砝码的指示旋钮来变换自10毫克-199.99克砝码以内所需重量值。 二、分析天平主要技术指标： 规格型号技术参数：秤盘尺寸 TG328A200g/0.1mg80mm TG328B200g/0.1mg80mm1 TG628A200g/1mg80mm 三、分析天平性能特点： 1.TG系列双盘机械天平，采用空气阻尼，光电读数，具有称量准确，读数简便，价格实惠。 2.专供实验室，学校，工矿等作精密称量分析用。 四、天平的安装方法： a．安装前的准备工作安装前的准备工作安装前的准备工作安装前的准备工作 1.安装选择：天平必须放在牢固的台上，不准有震动，气流存在，室内气温要求干燥明亮，温度最好保持在20℃±2℃左右，避免阳光晒射单面受热和气温潮湿，反之影响天平的灵敏度和正确性。
古代汉语练习题 词类活用
古代汉语练习（词类活用） 班级：姓名：学号： 一、简答： 1、什么是古代汉语的词类活用？古代汉语中的词类活用有哪几种？ 2、怎样区别使动用法和意动用法？试举例说明。并说明如何翻译。 3、试说明名词做状语主要有哪几种情况。 4、名词、形容词用作动词的情况主要有哪些？应该如何辨认？ 二、多项选择题（在每小题的四个备选答案中，选出二个至四个正确的答案，并将其号码分别填在题干后的括号内，多选、少选、错选均无分。每小题1分，共5分） 1．下列各句中加着重号的词，属于词类活用的是（） A．斩一首者爵一级B．能富贵将军者，上也 C．曹人凶俱，为其所得者棺而出之 D．夫鼠，昼伏夜动，不穴于寝庙，畏人故也 2．下列各句中加着重号的词属于名词作状语的是（） A．裂裳衣疮，手往善药 B．其经承子厚口讲指画为文词者，悉有法度可观 C．范增数目项王D．诸侯宾至 3．下列各句含宾语前置现象的是（） A．姜氏何厌之有B．楚君之惠，末之敢忘 C．除君之恶，唯力是视D．昭王南征而不复，寡人是问 4．对下列各句中加着重号的词组分析错误的是（） A．子重使太宰伯州犁待于王后（动宾）B．将塞井夷灶而为行也（连动） C．臣之壮也犹不如人（主谓）D．以勇力之所加而治智能之官（偏正） 5．下列句子中有使动用法的是（） A．秋九月，晋侯饮赵盾酒，伏甲将攻之 B．是时万石君奋为汉王中涓，受手谒，人见平
C．见灵辄饿，问其病，曰：“不食三日矣。”食之，舍其半 D．仓廪实而知礼节，衣食足而知荣辱 四、指出并具体说明下列文句中的词类活用现象： 1．秦数败赵军，赵军固壁不战。（秦与赵兵相距长平） 2．赵王不听，遂将之。（秦与赵兵相距长平） 3．身所奉饭饮而进食者以十数，所友者以百数。（秦与赵兵相距长平） 4．括军败，数十万之众遂降秦，秦悉阬之。（秦与赵兵相距长平） 5．信数与萧何语，何奇之。（韩信拜将） 6．王必欲长王汉中，无所事信。（韩信拜将） 7．吾亦欲东耳，安能郁郁久居此乎？（韩信拜将） 8．何闻信亡，不及以闻，自追之。（韩信拜将） 9．今大王举而东，三秦可传檄而定也。（韩信拜将） 10．遇有以梦得事白上者，梦得于是改刺连州。（柳子厚墓志铭） 11．自子厚之斥，遵从而家焉，逮其死不去。（柳子厚墓志铭） 12．以如司农治事堂，栖之梁木上。（段太尉逸事状） 13．踔厉风发，率常屈其座人。（柳子厚墓志铭） 14．晞一营大噪，尽甲。（段太尉逸事状） 15．即自取水洗去血，裂裳衣疮，手注善药。（段太尉逸事状） 16．黄罔之地多竹，大者如椽。竹工破之，刳去其节，用代陶瓦。（黄冈竹楼记）17．晋灵公不君。厚敛以彫墙。（晋灵公不君） 18．既而与为公介，倒戟以御公徒而免之。（晋灵公不君） 19．盛服将朝，尚早，坐而假寐。（晋灵公不君） 20．晋侯饮赵盾酒，伏甲将攻之。（晋灵公不君） 五、说明下列文句中的词类活用现象，并将全文译为现代汉语：
精神分裂症的发病原因是什么
精神分裂症的发病原因是什么 精神分裂症是一种精神病，对于我们的影响是很大的，如果不幸患上就要及时做好治疗，不然后果会很严重，无法进行正常的工作和生活，是一件很尴尬的事情。因此为了避免患上这样的疾病，我们就要做好预防，今天我们就请广州协佳的专家张可斌来介绍一下精神分裂症的发病原因。 精神分裂症是严重影响人们身体健康的一种疾病，这种疾病会让我们整体看起来不正常，会出现胡言乱语的情况，甚至还会出现幻想幻听，可见精神分裂症这种病的危害程度。 (1)精神刺激：人的心理与社会因素密切相关，个人与社会环境不相适应，就产生了精神刺激，精神刺激导致大脑功能紊乱，出现精神障碍。不管是令人愉快的良性刺激，还是使人痛苦的恶性刺激，超过一定的限度都会对人的心理造成影响。 (2)遗传因素：精神病中如精神分裂症、情感性精神障碍，家族中精神病的患病率明显高于一般普通人群，而且血缘关系愈近，发病机会愈高。此外，精神发育迟滞、癫痫性精神障碍的遗传性在发病因素中也占相当的比重。这也是精神病的病因之一。 (3)自身：在同样的环境中，承受同样的精神刺激，那些心理素质差、对精神刺激耐受力低的人易发病。通常情况下，性格内向、心胸狭窄、过分自尊的人，不与人交往、孤僻懒散的人受挫折后容易出现精神异常。 (4)躯体因素：感染、中毒、颅脑外伤、肿瘤、内分泌、代谢及营养障碍等均可导致精神障碍，。但应注意，精神障碍伴有的躯体因素，并不完全与精神症状直接相关，有些是由躯体因素直接引起的，有些则是以躯体因素只作为一种诱因而存在。 孕期感染。如果在怀孕期间，孕妇感染了某种病毒，病毒也传染给了胎儿的话，那么，胎儿出生长大后患上精神分裂症的可能性是极其的大。所以怀孕中的女性朋友要注意卫生，尽量不要接触病毒源。 上述就是关于精神分裂症的发病原因，想必大家都已经知道了吧。患上精神分裂症之后，大家也不必过于伤心，现在我国的医疗水平是足以让大家快速恢复过来的，所以说一定要保持良好的情绪。
基于单片机的自动存包系统设计
基于单片机的自动存包系统设计 摘要 近年来，随着生活水平的提高，人们对于社会消费品的质量和数量的要求也在逐渐增加。为了更好的为广大顾客服务，在一些商场、影院、超市等公共场合通常设置有自动存包柜，本次便是针对这一现象进行设计。 本文详细介绍了国内自动存包控制系统的发展现状，发展中所面临的问题。并详细介绍了本系统采用的AT89S52单片机做控制器，可以同时管理四个存包柜。柜门锁是由继电器控制，当顾客需要存包的时候，可以自行到存包柜前按“开门”键，需要顾客向光学指纹识别系统输入个指纹，然后通过继电器进行开门（用亮灯表示），顾客即可存包，并需将柜门关上。当顾客需要取包时，要将只要将之前输入的指纹放置于指纹识别器前方，指纹识别器采集到指纹信息输出相应的高低电平信号传给单片机，系统比较密码一致后，发出开箱信号至继电器将柜门打开，顾客即可将包取出。它具有功能实用、操作简便、安全可靠、抗干扰性强等特点。 关键词：自动存包柜,单片机,指纹识别器
李少鹏：基于单片机的自动存包系统设计 Based on single chip microcomputer automatic package design Abstract In recent years, with the improvement of living standards, people for social consumer goo ds quality and quantity requirements are to increase gradually. In order to better service for the g eneral customers, in some stores, movie theaters, supermarkets public Settings are to be put auto matically usually bag ark, it is functional practical, simple operation, safe and reliable, anti-jamm ing strong sexual characteristics. Domestic deposit automatic control system are introduced in detail in this paper the development of the status quo, problems faced in the development of. And introduces in detail the system adopts single chip microcomputer controller, can simultaneously manage multiple pack ark. Cupboard door lock controlled by relay, when customers need to save package, will be allowed to save package before the ark according to the "open" button, need customer to the system input fingerprint, and then through the relay to open the door (with lighting), customers can save package, and the cupboard door must be closed. When customers need to pick up package, as long as before the input fingerprint should be placed on the fingerprint recognizer, fingerprint recognizer collecting to the fingerprint information and output the corresponding high and low level signal to the microcontroller, the system is password consistent, signal out of the box to the relay Key words: Automatic Storage Bag, Microcontroller, Fingerprint recognizer。
电子精密天平秤的使用方法及注意事项(正式版)
电子精密天平秤的使用方法及注意事项 刘维彬电子精密天平秤是定量分析工作中不可缺少的重要仪器，充分了解仪器性能及熟练掌握其使用方法，是获得可靠分析结果的保证。精密天平的种类很多，有普通精密天平、半自动/全自动加码电光投影阻尼精密天平及电子精密天平等。下面就电子精密天平的使用方法及注意事项做一介绍。 操作方法: 1．检查并调整天平至水平位置。 2．事先检查电源电压是否匹配（必要时配置稳压器），按仪器要求通电预热至所需时间。 3．预热足够时间后打开天平开关，天平则自动进行灵敏度及零点调节。待稳定标志显示后，可进行正式称量。 4．称量时将洁净称量瓶或称量纸置于称盘上，关上侧门，轻按一下去皮键，天平将自动校对零点，然后逐渐加入待称物质，直到所需重量为止。 5．称量结束应及时除去称量瓶（纸），关上侧门，切断电源，并做好使用情况登记。 注意事项: 1.天平应放置在牢固平稳水泥台或木台上，室内要求清洁、干燥及较恒定的温度，同时应避免光线直接照射到天平上。 2.称量时应从侧门取放物质，读数时应关闭箱门以免空气流动引起天平摆动。前门仅在检修或清除残留物质时使用。 3.电子精密天平若长时间不使用，则应定时通电预热，每周一次，每次预热
2h，以确保仪器始终处于良好使用状态。 4.天平箱内应放置吸潮剂(如硅胶)，当吸潮剂吸水变色，应立即高温烘烤更换，以确保吸湿性能。 5.挥发性、腐蚀性、强酸强碱类物质应盛于带盖称量瓶内称量，防止腐蚀天平。 6.称量重量不得过天平的最大载荷。 7.经常对电子天平进行自校或定期外校，保证其处于最佳状态。 8.天平发生故障，不得擅自修理，应立即报告测试中心质量负责人。 9.天平放妥后不宜经常搬动。必须搬动时，移动天平位置后，应由市计量部门校正计量合格后，方可使用。
初中所学文言文中的五类常见词类活用现象
初中所学文言文中的五类常见词类活用现象
古代汉语中的词类活用现象 五种类型：名词用作动词 动词、形容词、名词的使动用法 形容词、名词的意动用法 名词用作状语 动词用作状语 （一）名词用如动词 古代汉语名词可以用如动词的现象相当普遍。如： 从左右，皆肘．之。（左传成公二年） 晋灵公不君．。（左传宣公二年） 孟尝君怪其疾也，衣冠 ．．而见之。（战国策·齐策四） 马童面．值，指王翳曰：“此项王也。”（史记·项羽本纪） 夫子式．而听之。（礼记·檀弓下） 曹子手．剑而从之。（公羊传庄公十三年） 假舟楫者，非能水．也，而绝江河。（荀子·劝学） 左右欲刃．相如。（史记·廉颇蔺相如列传） 秦师遂东．。（左传僖公三十二年） 汉败楚，楚以故不能过荥阳而西．。（史记·项羽本纪） 以上所举的例子可以分为两类：前八个例子是普通名词用如动词，后两个例子是方位名词用如动词。 名词用作动词是由上下文决定的。我们鉴别某一个名词是不是用如动词，须要从整个意思来考虑，同时还要注意它在句中的地位，以及它前后有哪些词类的词和它相结合，跟他构成什么样的句法关系。一般情况有如下四种：
①代词前面的名词用如动词（肘之、面之），因为代词不受名词修饰； ②副词尤其是否定副词后面的名词用如动词（“遂东”、“不君”）； ③能愿动词后面的名词也用如动词（“能水”、“欲刃”）； ④句中所确定的宾语前面的名词用如动词（“脯鄂侯”“手剑”） （二）动词、形容词、名词的使动用法 一、动词的使动用法。 定义：主语所代表的人物并不施行这个动词所表示的动作，而是使宾语所代表的人或事物施行这个动作。例如：《左传隐公元年》：“庄公寤生，惊姜氏。”这不是说庄公本人吃惊，而是说庄公使姜氏吃惊。 在古代汉语里，不及物动词常常有使动用法。不及物动词本来不带宾语，当它带有宾语时，则一定作为使动用法在使用。如： 焉用亡．郑以陪邻？《左传僖公三十年》 晋人归．楚公子榖臣与连尹襄老之尸于楚，以求知罃。（左传成公三年） 大车无輗，小车无杌，其何以行．之哉？《论语·为政》 小子鸣．鼓而攻之可也。《论语·先进》 求也退，故进．之；由也兼人，故退．之。《论语·先进》 故远人不服，则修文德以来．之。《论语·季氏》 有时候不及物动词的后面虽然不带宾语，但是从上下文的意思看，仍是使动用法。例如《论语·季氏》：“远人不服而不能来也”这个“来”字是使远人来的意思。 古代汉语及物动词用如使动的情况比较少见。及物动词本来带有宾语，在形式上和使动用法没有什么区别，区别只在意义上。使动的宾语不是动作的接受者，而是主语所代表的人物使它具有这种动作。例如《孟子·梁惠王上》“朝秦楚”，不食齐宣王朝见秦楚之君，相反的，是齐宣王是秦楚之君朝见自己。 下面各句中的及物动词是使动用法： 问其病，曰：“不食三日矣。”食．之。《左传·宣公二年》
精神分裂症的病因是什么
精神分裂症的病因是什么 精神分裂症是一种精神方面的疾病，青壮年发生的概率高，一般 在16～40岁间，没有正常器官的疾病出现，为一种功能性精神病。 精神分裂症大部分的患者是由于在日常的生活和工作当中受到的压力 过大，而患者没有一个良好的疏导的方式所导致。患者在出现该情况 不仅影响本人的正常社会生活，且对家庭和社会也造成很严重的影响。 精神分裂症常见的致病因素： 1、环境因素：工作环境比如经济水平低低收入人群、无职业的人群中，精神分裂症的患病率明显高于经济水平高的职业人群的患病率。还有实际的生活环境生活中的不如意不开心也会诱发该病。 2、心理因素：生活工作中的不开心不满意，导致情绪上的失控，心里长期受到压抑没有办法和没有正确的途径去发泄，如恋爱失败， 婚姻破裂，学习、工作中不愉快都会成为本病的原因。 3、遗传因素：家族中长辈或者亲属中曾经有过这样的病人，后代会出现精神分裂症的机会比正常人要高。 4、精神影响：人的心里与社会要各个方面都有着不可缺少的联系，对社会环境不适应，自己无法融入到社会中去，自己与社会环境不相
适应，精神和心情就会受到一定的影响，大脑控制着人的精神世界， 有可能促发精神分裂症。 5、身体方面：细菌感染、出现中毒情况、大脑外伤、肿瘤、身体的代谢及营养不良等均可能导致使精神分裂症，身体受到外界环境的 影响受到一定程度的伤害，心里受到打击，无法承受伤害造成的痛苦，可能会出现精神的问题。 对于精神分裂症一定要配合治疗，接受全面正确的治疗，最好的 疗法就是中医疗法加心理疗法。早发现并及时治疗并且科学合理的治疗，不要相信迷信，要去正规的医院接受合理的治疗，接受正确的治 疗按照医生的要求对症下药，配合医生和家人，给病人创造一个良好 的治疗环境，对于该病的康复和痊愈会起到意想不到的效果。
自动存包柜的设计与仿真
自动存包柜的设计与仿真 摘要 本课题是基于单片机的自动存包柜设计。自动存包柜是新一代的存包柜，具有功能实用、操作简单、管理方便、安全可靠等特点，能够更好的服务于不同市场的广大群众，使用者可以根据简明清晰的操作说明自行完成存包取包工作。本系统由MCS-51单片机构成核心控制系统，整个系统由主控部分、键盘显示控制部分、执行部分三部分组成，通过随机密码的产生和核对完成自动存包取包过程。本设计中各元器件便于安装且操作简单，能基本实现存包取包功能。 关键词：自动存包柜；单片机；随机密码
Design and Simulation of Automatic Lockers ABSTRACT This topic is microcontroller-based automatic lockers.Automatic lockers is a new generation of lockers, with a practical, simple operation, easy management, safe and reliable, able to better serve the broad masses of the different markets, users are based on a clear and concise instructions to complete the deposit bags to take the package. The system consists of MCS-51 microcontroller core control system, the entire system from the main section, the keyboard display control part of the implementation of some of the three-part composition, random password generation and check completed automatically save the package to take the package process. Various components of this design is easy to install and easy to operate, can basically save the package to take package function. Key words :Automatic lockers; microcontroller; random password
大件寄存管理流程
大件寄存程序 一、大件寄存流程 1、当顾客提出大件寄存需求时，服务中心员工应先查看物品大小，如果可自助存包柜寄存的物品，请顾客优先使用自动存包柜，并提醒顾客不要寄放贵重物品。 2、如果物品过大，自助存包柜寄存不下的物品或顾客坚持要求寄存服务台的商品，服务台员工应与顾客确认是否有贵重物品，如有，请顾客将贵重物品取出自行保管，其它物品则办理寄存手续。 3、员工将顾客物品放置在指定区域，挂好存包牌，将另一块号码牌交予顾客，并告知顾客于当天营业结束前凭此号码牌领取寄存物品，请顾客妥善保管。 4、当顾客前来领取寄存物品时，服务中心员工请顾客出示号码牌，员工收回号码牌，根据号码牌查找物品上对应的号码牌，核对无误后交还顾客物品。存包牌两块互绑在一起放回存放盒内。 5、若核对号码牌时发现物品与存放时不符或丢失,应立即告知客服主管或值班经理前来处理(具体处理方法请参照客诉处理办法) 6、如顾客遗失号码牌时，服务中心员工处理程序： A、通知广播寻找遗失的号码牌。 B、询问顾客具体号码牌号或所寄存物品形状、颜色、特征。 C、询问顾客是否有任何的证件，并请顾客提供有效证件证明所寄存物品确为其所属。 D、请顾客在《寄存物异常登记表》上留下地址、电话、证件号等信息。 E、向顾客适当收取号码牌工本费并在退换货收银机做销售操作，将寄存物品归还于顾客。 7、每日营业结束后盘点号码牌，每块号码牌要求一个号码两块互绑在一起，如已不齐全的号码牌则不再使用。号码牌需定期全部更换。 二、大件寄存顾客告示 1、请您在存包后主动索取存包牌，并保存好自己的存包牌，不要遗失。 2、现金、手机、照相机、存折等超过200元的贵重物品及有效证件如身份证、户口本等物品请随身携带，请勿寄存，由此造成的损失，自行承担。 3、当您领取所存物品时，请出示存包牌，并当面点清您所存的物品，离柜本商场不再负责。 4、如存包牌丢失，应及时持有效证件领取所存物品，并付工本费5元，未及时挂失，导致所存物品丢失，不予负责。
词类活用例子
文言实词词类活用 活用为一般动词 （一）名词活用为一般动词 1.两个名词连用，既不是并列关系，又不是修饰关系，便是动宾或主谓，其中一个必然活用为动词。 a .有一老父，衣褐，至良所。 b.籍吏民，封府库。 c.我有嘉宾，鼓瑟吹笙。 d.冬雷震震夏雨雪。 2.名词后紧跟代词，该名词活用为动词。 a.驴不胜怒，蹄之。 b.以其乃华山之阳名之。 c.名余曰正则兮。 3.名词放在副词后，便活用为动词。 a.日将暮，取儿槁葬。 b.太子及宾客知其事者，皆白衣冠以送之。 c.从弟子女十人所，皆衣缯单衣，立大巫后。 4.名词放在“能”“可”“足”“欲”等呢过愿动词后，便活用为动词。 a.假舟楫者，非能水也。 b.云青青兮欲雨。 c.其力尚足以入，火尚足以明。 d.子谓公冶长：“可妻也。” 5.名词带介宾结构做补语，这个名词活用为动词。 a.晋军（于）函陵，秦军（于）氾南。 b.唐浮图慧褒始舍于其址。 6.名词用“而”同动词或动宾词组连接时，活用为动词。 a.三代不同礼而王，五霸不同法而霸。 7.名词在“所”“者”结构中便活用为动词。 a.置人所罾鱼腹中。
a.是以，令吏人完客所馆。 形容词活用为一般动词 1.形容词用在“所”字之后，便活用为动词。 故俗之所贵，主之所贱；吏之所卑，法之所尊也。 （认为宝贵、认为低贱、认为卑下、认为高贵） 2.形容词在能愿动词后，活用为动词。 问其深，则其好游者不能穷也。（走到尽头） 3.形容词在“之”“我”能代词前，活用为动词。 稍出近之。（靠近） 4.形容词后带介宾结构做补语，它活用为动词。 令尹子兰……率使上官大夫短屈原于顷襄王。 （诋毁） 数词活用做一般动词 六王毕，四海一。（统一） 名词做状语 一、普通名词作状语 1.表比喻 a.嫂蛇行匍匐。 b.狐鸣呼曰。 c .赢粮而景从。 d .天下云集响应。 e.常以身翼蔽沛公。 f.一狼径去，其一犬坐于前。 2.表对人的态度 a.君为我呼入，吾得兄事之。 b.人人皆得以隶使之。 3.表动作行为的处所 a.夫以秦王之威，相如廷叱之，辱其群臣廷：在朝廷上 b.童子隅坐而执烛. 隅：在墙角 4.表动作行为的工具、凭借、方式
精神分裂症应该怎么治疗
精神分裂症应该怎么治疗 1、坚持服药治疗 服药治疗是最有效的预防复发措施临床大量统计资料表明，大多数精神分裂症的复发与自行停药有关。坚持维持量服药的病人复发率为40%。而没坚持维持量服药者复发率高达80%。因此，病人和家属要高度重视维持治疗。 2、及时发现复发的先兆，及时处理 精神分裂症的复发是有先兆的，只要及时发现，及时调整药物和剂量，一般都能防止复发，常见的复发先兆为：病人无原因出现睡眠不好、懒散、不愿起床、发呆发愣、情绪不稳、无故发脾气、烦躁易怒、胡思乱想、说话离谱，或病中的想法又露头等。这时就应该及时就医，调整治疗病情波动时的及时处理可免于疾病的复发。 3、坚持定期门诊复查 一定要坚持定期到门诊复查，使医生连续地、动态地了解病情，使病人经常处于精神科医生的医疗监护之下，及时根据病情变化调整药量。通过复查也可使端正人及时得到咨询和心理治疗解除病人在生活、工作和药物治疗中的各种困惑，这对预防精神分裂症的复发也起着重要作用。 4、减少诱发因素 家属及周围人要充分认识到精神分裂症病人病后精神状态的薄弱性，帮助安排好日常的生活、工作、学习。经常与病人谈心，帮助病人正确对待疾病，正确对待现实生活，帮助病人提高心理承受能力，学会对待应激事件的方法，鼓励病人增强信心，指导病人充实生活，使病人在没有心理压力和精神困扰的环境中生活。 首先是性格上的改变，塬本活泼开朗爱玩的人，突然变得沉默寡言，独自发呆，不与人交往，爱干净的人也变的不注意卫生、生活
懒散、纪律松弛、做事注意力不集中，总是和患病之前的性格完全 相悖。 再者就是语言表达异常，在谈话中说一些无关的谈话内容，使人无法理解。连最简单的话语都无法准确称述，与之谈话完全感觉不 到重心。 第三个就是行为的异常，行为怪异让人无法理解，喜欢独处、不适意的追逐异性，不知廉耻，自语自笑、生活懒散、时常发呆、蒙 头大睡、四处乱跑，夜不归宿等。 还有情感上的变化，失去了以往的热情，开始变的冷淡、对亲人不关心、和友人疏远，对周围事情不感兴趣，一点消失都可大动干戈。 最后就是敏感多疑，对任何事情比较敏感，精神分裂症患者，总认为有人针对自己。甚至有时认为有人要害自己，从而不吃不喝。 但是也有的会出现难以入眠、容易被惊醒或睡眠不深，整晚做恶梦或者长睡不醒的现象。这些都有可能是患上了精神分裂症。 1.加强心理护理 心理护理是家庭护理中的重要方面，由于社会上普遍存在对精神病人的歧视和偏见，给病人造成很大的精神压力，常表现为自卑、 抑郁、绝望等，有的病人会因无法承受压力而自杀。家属应多给予 些爱心和理解，满足其心理需求，尽力消除病人的悲观情绪。病人 生活在家庭中，与亲人朝夕相处，接触密切，家属便于对病人的情感、行为进行细致的观察，病人的思想活动也易于向家属暴露。家 属应掌握适当的心理护理方法，随时对病人进行启发与帮助，启发 病人对病态的认识，帮助他们树立自信，以积极的心态好地回归社会。 2.重视服药的依从性 精神分裂症病人家庭护理的关键就在于要让病人按时按量吃药维持治疗。如果不按时服药，精神病尤其是精神分裂症的复发率很高。精神病人在医院经过一系统的治疗痊愈后，一般需要维持2~3年的
词类活用例子
词类活用例子 Company Document number：WUUT-WUUY-WBBGB-BWYTT-1982GT
文言实词词类活用 活用为一般动词 （一）名词活用为一般动词 1.两个名词连用，既不是并列关系，又不是修饰关系，便是动宾或主谓，其中一个必然活用为动词。 a .有一老父，衣褐，至良所。 b.籍吏民，封府库。 c.我有嘉宾，鼓瑟吹笙。 d.冬雷震震夏雨雪。 2.名词后紧跟代词，该名词活用为动词。 a.驴不胜怒，蹄之。 b.以其乃华山之阳名之。 c.名余曰正则兮。 3.名词放在副词后，便活用为动词。 a.日将暮，取儿槁葬。 b.太子及宾客知其事者，皆白衣冠以送之。 c.从弟子女十人所，皆衣缯单衣，立大巫后。 4.名词放在“能”“可”“足”“欲”等呢过愿动词后，便活用为动词。 a.假舟楫者，非能水也。 b.云青青兮欲雨。 c.其力尚足以入，火尚足以明。 d.子谓公冶长：“可妻也。”
5.名词带介宾结构做补语，这个名词活用为动词。 a.晋军（于）函陵，秦军（于）泛南。 b.唐浮图慧褒始舍于其址。 6.名词用“而”同动词或动宾词组连接时，活用为动词。 a.三代不同礼而王，五霸不同法而霸。 7.名词在“所”“者”结构中便活用为动词。 a.置人所罾鱼腹中。 a.是以，令吏人完客所馆。 形容词活用为一般动词 1.形容词用在“所”字之后，便活用为动词。 故俗之所贵，主之所贱；吏之所卑，法之所尊也。（认为宝贵、认为低贱、认为卑下、认为高贵）2.形容词在能愿动词后，活用为动词。 问其深，则其好游者不能穷也。（走到尽头） 3.形容词在“之”“我”能代词前，活用为动词。 稍出近之。（靠近） 4.形容词后带介宾结构做补语，它活用为动词。 令尹子兰……率使上官大夫短屈原于顷襄王。 （诋毁） 数词活用做一般动词 六王毕，四海一。（统一） 名词做状语