Copyright Notice: This article can be reproduced, please be sure to indicate the original source and author information in hyperlink when reprint. Original Source: http://www.aiview.com/focus/cms_thinking.htm Author: Zhang Yang Alex_doesAThotmail.com Last updated: 2003-9-17
table of Contents
Text reference resources
text
Today, with highly developed information technology, more and more people realize that they are facing different types of management, and growing digital resources. When these resources accumulate to a certain extent, every person will ask for an identical Question, it is also a problem I have been thinking - how do you manage the large amount of resources collected very good? The file types of these resources are varied, such as HTML, PDF, TXT, JPG, ZIP, EXE, TAR, GZ, etc., and various types of electronic books, software code, pictures, maybe you like. Literary works and movies, even your important emails a year ago, in short, you are difficult to use a classification method to organize them very well, and some of them have a lot of things, such as your travel Photo, you can put it in the top-level name of the picture - Of course, this is more complicated, so this is clear. But you probably hope to read this travel time. It can also be very convenient to see it, so I have put this picture in my document -> Tourism -> The demand in the picture directory, saving two copies, there is a problem. Very obvious, traditional documents The tree organization of the system is no longer able to meet the needs of your convenient access to the file. You need a better file access method, you can find this photo through the picture class, you can also read you half a year. Journey to see it, maybe you will also expect to have more ways. So you will find a way to classify these resources, and make a cross link to the upper, even if you don't care how they store, how Implementation, just give logical relationships, but you may find that this is not an easy thing.
So you will expect a content management system to help deal with this all things, my thoughts are also starting here, and try to give a clearer definition.
First, I just treat it only as a classification management system, not responsible for the real data storage, which should be made by another system, such as (object / relationship) database, file system, etc. Between these two systems should be very Good collaboration, communication, but should not be coupling too tight, you need a clear interface.
Secondly, I think this classification management system should provide such a function, a small text handler can go out to analyze and process the currently browsed text - such as a web page, or a batch of analytical processing a directory, or one Word documentation, put this information according to my requirements and save it, can be called "import". This feature does not belong to the core of the "Classification Management System", which is just an imported plugin for different types of data sources. Such an interface should be left over this system, so that different people can develop a plug-in for different data types. However, if you want to make it easier to implement data extraction, more dependent on source data.
About the data of the data, abstraction of an object's properties, abstract from its various properties, to distinguish other human assignment social attributes. For example, the gender of people is relatively stable, which is inherent attributes, others such as titles. , Work units, home address, etc. are some social properties. This is itself a deep enough topic that can be discussed.
The bottom of the underlying data storage should also be user-controllable. For example, an article you see on the web page may be relative to one or more files in the backend, and the arrangement of these files is not unordered, you should be The storage structure of these files is speculated from the representation layer, which is convenient to access. This is very convenient for backed up or small range operations, although you have other means of this system, but this is your most basic right. At the end of this system, I want to provide a web-based data representation to the user.
Now you can imagine this system as a bridge, one side is basic data, one side is the beneficion group of these data. This bridge generates an index according to the basic data, while on the other side, according to the benefit group according to the data, View. Basic data can be considered as the matrix, classification is relatively stable. The view can be multiple, the view is also reflected in some category, but the basis of this classification is the usage habit of data benefit groups, in some cases Next, the viewing method of the view may be consistent with the classification of the basic data. This, there is a mapping relationship between the category of the basic data and the category of the user view, and the view provided to the user may be a subset of all basic data, also It can be all. This mapping relationship is this bridge, which is exactly a set of bridges.
For basic data, it can be a new data collection, or an existing, classified data collection. Our bridge does not try to change the original data classification method and classification standard, first, according to the present Some data content provides a suitable view for the data benefit group and makes a mapping relationship between the classification and views of the basic data.
The establishment and selection of user views, and how to build mappings with basic data is completed by experts who are familiar with these data, this system does not replace you. It is better to give you a bookshelf, just this However, how to organize your book, and what location should be placed, this is your business. Of course, I will make the bookshelf more comfortable, more convenient for your reading and use.
Compared to the original idea - this system manages all resources, now some different. We think it is proprietary, such as one such system only manages book resources, and another such system specifically manages audiovisual resources, possibly There are other special management for the original code for the computer, and these multiple different, integrated, scalable systems are put together, and achieve the initial management of all resources. You can use each The management of resources is regarded as a component, and we can easily add a new component to manage new resource types. This requires that these components can be collaborative, and to design well, it is too loose, Can't reach the goal of collaboration; it will become very complicated, and the meaning of the assembly is too tight, and it will become more complicated.
The following content will be deeply in-depth and will design some architectural considerations.
This system is based on an existing, classified data collection. If there is not yet, we can help you build one. We use the data in this collection only to the data itself, does not contain data Other additional properties, even in this collection contains these additional descriptions, we will not pay. Because this will cause our system and existing interface interface complications, this is for us to work in most existing existing existing The desire to data collection is contrary.
We put these basic data attributes in our system's internal storage, which is the necessary value, this is the value of this system and the premise of work. This will bring two work.
First, the import problem of the original basic data related information, which contains the access location, access method, related attributes of the original data. Relying on the original amount of data, this workload may be very huge, there will be some required Manual operation and manual intervention, most of the work hopes to be done by one or more import scripts, such as using Perl.
Second, there may be such a case, the original data has some unique properties that the user rely on, and our system is happens that there is no description and processing of this attribute. So we face it in the system. Additional work for processing.
For the second question, we should try to avoid it, that is, when we design the components of certain resource types, we must cover the needs of users to filter these features by configuring different users. In this case In the case, if there is a special user needs, I think users will not mind this extra development cost, which is he should bear, the interests you can get from it.
In addition, we can continue to abstract, expand some of the applicable scope of basic data collection, exceeding the category of digital resources, such as a library that stores books in paper. So, whether our system should also be suitable It works well, in addition to the operations of those online on basic data itself, such as revision, increase, etc., but we can still operate online attributes, such as book name and borrowing status, proceed Procedures, etc. In this case, the location of this book saved inside our system should be his bookshelf number, not an electronic link.
/ User View One - User Group
Electronic Book Resources Electronic Book Management Components - - User View II - - User Group 2
// / user view three - - User group three
//
// / User View One /
Digital Audiovisual Resource Digital Audiovisual Management Component - - User View II - - User Group
// / User view three /
//
//
Traditional Library Traditional Book Management Components - - User View - - User Group
Basic Data Source Resource Management Components
----------------------------------------------------------------- -----------------------------
| | | | | | | | |
| File System | | Import | XML | | Import | User |
| Object Relationship DB | <=> | Script | Or | Logic Floor | Plugin | View |
| LDAP Data Source | | => | Relationship DB | | <= | |
| Non-Digital Resources | | | | | | |
----------------------------------------------------------------- -----------------------------
Recently, due to the needs of the project, I went to Fang Zheng Institute to read the functional demonstration of Founder Boss 3.0. I feel that Boss's function is relatively perfect, and many types of data resources have been managed. For example, there is already a single document, composite document, composite asset and other concepts; Features. Bis needs to work on a large database, all managed data, including stream of stream of large data, is stored in large object mode, not in the form of file system. That is, there is no clear interface with existing classification resources.
Reference resource
Thanks to Gong Jian, Gong Jian, provided by the beneficial ideas and inspiration! CMS REVIEW, http://www.cmsreview.com cms Watch http://www.cmswatch.com
