China DOS Union

The idea of organizing classic posts in the Batch Processing Room was proposed a long time ago, but at that time, there was not enough time, resulting in the suspension of this matter. In early January, under the continuous updates of站长wengier and the hard tests by qzwqzw and others, the development work of the DOS interface in the forum went very smoothly, which greatly facilitated the extraction of post title records, thus creating convenient conditions for establishing the index of classic posts in the Batch Processing Room.

From the end of January, I started to organize the index of classic posts in the Batch Processing Room. However, I was occupied with various things in the middle, and the organizing work was interrupted for some time. I just continued to do it the day before yesterday, and finally finished the first draft today. I put it out for everyone to discuss and see how to further improve it.

Let me talk about my inclusion criteria:
① Posts that make in-depth analyses and discussions on a certain issue;
② Posts with high concern, strong practicality/generality;
③ Posts with relatively high skills or unique ideas/insights;
④ Posts that provide different ideas/different script tool writing codes.

Since it is only a first draft, two versions of the organizing results are provided: Considering the convenience of organizing, "Classified Organization Preliminary (Detailed).txt" only retains key contents such as the creation time, ID, number of replies, page views, and titles of the posts, without adding the forum links, but it can be accessed by the way of http://www.cn-dos.net/forum/viewthread.php?tid=ID number; Considering the convenience of viewing the post contents, "Classified Organization Preliminary (With Links Version).txt" only provides the links and titles. Since it is a first draft, there must be many deficiencies, for example, the classification is not reasonable enough, and there are places that can be added/deleted/adjusted; for example, there may be content repetitions in similar posts, and we can also discard the irrelevant and keep the essential... Please give more suggestions to make this post index a fine product in the Batch Processing Room.

At present, the preliminary ideas are:
① Rationalize the index classification more;
② Try to be as concise as possible, and only select representative similar posts to post;
③ The posted index posts will be made like this: Essence Post Index of the Q&A Room in the DOS Union Forum (2005.08.10), however, it will be posted in the form of classification, and distinguished by means of adding colors, changing font sizes, etc.

Thanks to ccwan for the suggestions on classification; thanks to jmz573515 for the suggestions on the version contents.

Super top!!! It's really great～: ) ) )

Originally posted by redtek at 2007-2-6 13:07:
Super top!!! It's really great~:)

　　Looking forward to your suggestions.

Brother namejm is always so modest～:)

Looking forward to the brilliant index post of the elite threads in the DOS Batch Processing Room of the DOS Union Forum hanging brightly in the "Important Topics" section of the forum～:)

Damn you!
Damn you to the core!

In order to enable the index to be released as soon as possible and facilitate everyone's learning, I hope everyone will not stand by idly.

Very good!
It is suggested that when publishing the index, the posting time should be attached, like the post with tid=13226. Brother namejm is currently organizing and keeping the last reply time, which seems to be of little significance.

Originally posted by Wu Nai He at 2007-2-7 13:35:
It is suggested that when releasing the index, the posting time should be attached, like the post with tid=13226. Brother namejm is currently sorting and keeping the last reply time, which seems to be of little significance.

　　As mentioned in the top floor, the format for releasing the index refers to the format of the sticky index thread in the Q&A room, which will include the posting time. And in the preliminary sorting result I released currently, what is kept is also the posting time, not the last reply time. Brother Wu Nai He might have misread.

[ Last edited by namejm on 2007-8-13 at 01:56 PM ]

namejm Brother, I'm sorry, I just took a quick look.

Classification is indeed a headache. If the classification is too broad, it will seem messy; if it's too detailed, it's hard to judge which category a post should go into. I think it's better to have more detailed classification. Now there's a lack of a systematic and reasonable classification table, and I hope everyone can discuss it together.

Actually, there's no need to be too particular about the classification method. Because many contents are impossible to classify. It may belong to both A and B, and be related to C at the same time. The key is to let people quickly find the content they need. So the significance of establishing an index is greater than establishing a classification. The most effective and commonly used in the index is undoubtedly the keyword index. And the establishment and selection of keywords are much simpler. You can extract high-frequency words from the content, or summarize keywords from the content.

-------------------------------------------------------

In addition, there is a preliminary idea -

Because the quality of keywords depends on the accuracy of positioning and the attention degree of users, so it can be considered to dynamically optimize the index with the user's choices. Continuously add new keywords and their indexes, and eliminate old keywords and their indexes. This process should be able to be automatically realized. The key is to adopt what kind of method to collect and feedback users' choice behaviors.

Originally posted by qzwqzw at 2007-2-8 05:27:
Therefore, we can consider dynamically optimizing the index according to user choices

Your proposal is very good. Implementing dynamic indexing will be very difficult for me because I don't know JavaScript and so on. Please also ask friends who know JavaScript to give some guidance on whether it is feasible.

The plan I considered is to make a CHM file, set up a micro Apache + PHP on this machine to receive the user-submitted keywords, category information, etc., and then uniformly collect each user's submission information, and finally make a CHM file with categories and indexes. In any case, this cannot achieve real dynamism and cannot be updated at any time with the user's organization.

If we go further, we can use scripts to generate the files required for CHM compilation, and let users compile the CHM file by themselves.

Originally posted by qzwqzw at 2007-2-7 16:27:
Actually, there is no need to be too particular about the classification method
Because many contents cannot be classified
It may belong to both A and B, and be related to C at the same time
The key is to let people quickly find the content they need
So the significance of establishing an index is greater than establishing a classification

　　Some contents are indeed difficult to be accurately classified, but I think the work of classification should still be done, because according to the preliminary sorting results, there are more than 400 classic post titles. If they are not placed in different categories but all put together, it will be very difficult to find the content you are interested in - in order to reduce the sorting intensity and difficulty, I only want to list the post titles for now, and users can jump to this post by clicking. What I am doing is a title list index, not a full-text keyword index.

　　Of course, if there is a way to make a keyword index, it will be a great blessing in the forum. Unfortunately, with my level, I can't do this high-difficulty project, and I can only count on others

I don't know if this can bring convenience to everyone