- The reason hierarchical classification falls apart is that it should update the classification method when something appears that doesnât fit the first classification method, but it doesnât do that, Scrapboxâs philosophy that âthatâs why all hierarchical classification is out of order and shouldnât be doneâ is going a bit too far, The plain âContainer Metaphorâ is easy for many people to catch on to, and I donât think itâs right that kintone should be âthere is no such thing as space, all posts are flatâ. If rearrange is systematically supported when the container metaphor is broken, then there is no problem doing hierarchical classification.
https://twitter.com/nishio/status/1334051841680699392
- I am beginning to feel that âsupport for reorganizing informationâ is needed, further developed from the discussion the other day, so I will write a series of posts about it.
- The story of âhierarchical classification will eventually fail,â and even if that were true, âwe should not do hierarchical classification at allâ is not true, and there is a logical leap here.
- There were two major causes for the breakdown of hierarchical classification. (1) the appearance of items that cannot be classified well using the classification method assumed in advance, and (2) discrepancies in classification methods when used by multiple people.
- If the former is also seen as âa discrepancy between past and new classification methods,â this is âa problem with a system that can only have one classification method.
- Because there is only one way to classify information, when information that does not fit well is found, it is necessary to âdiscard the past classification method and reclassify it using the new method,â which most people find troublesome and force information into the old classification method.
- Reclassificationâ is not done because the cost of âreclassificationâ is too high. If this is the case, why not design a system that facilitates reclassification, or a system that supports reclassification?
- For example, when âsomething you donât know where to putâ appears, it is often put in an appropriate place because the user feels pressure to âput it somewhere right away. To prevent this, we should prepare a âplace to put things you donât know where to putâ in advance.
- This is a common operation in comic book cafes. It is better to put them in the category of âunclassified itemsâ than to ignore the classification and put them in an appropriate place.
- Next, we need assurance that âreclassificationâ will not destroy anything. What is wrong with simple hierarchical classification by folders and files is that if they are moved to a different location after reorganization, they will not be found in their original location. The system should be designed to take care of this.
- For example, a system of container metaphors in which âtagsâ are virtual folders and items with those âtagsâ are shown in them.
- If you move file X from folder A to folder B, âX that should have been in A is gone!â However, even if tag B is added to file X with tag A, it will not lose its âvirtual state of being in Aâ. With this guarantee, you can add tags with peace of mind.
- The two tools I have given so far are not enough at all. For example, if one person puts X in A and another person moves it to B, saying âit is wrong that this is in A,â if X is gone from A, the person who thinks it is in A is lost, and if it is not gone, the expression of the intention âit is wrong that it is in Aâ is not possible.
- One solution is to add a âtrash boxâ to A that contains âdeleted or moved items from Aâ. Another solution is to search for âitems that were in A in the pastâ.
- Just as it is difficult to determine an appropriate classification method in advance for hierarchical classification, it is also difficult to determine appropriate tagging criteria in advance for tagging. As a result, a single tag may appear on 100 pages, making it difficult to list them.
- It is natural to think, âIf all occurrences of the word X are made into tags, it will be no different from the search results with X, so you should not do that.
- An analogy to software development is that it is difficult to properly design in advance what references what. What to do, then, is to ârefactor when inconveniences become apparent. Refactoring means ârewriting the current functionality so that it does not break.
This page is auto-translated from /nishio/ć´çăç´ăăă¨ăŽćŻć´ using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. Iâm very happy to spread my thought to non-Japanese readers.