Please finally shoot these tags!
I love cool articles, I need them like air. If some day comes, when I cannot read a beautiful article, the world will stop for me. When I see a good idea, I immediately write a letter to the author and we have a heated discussion. But here's the trap - tags drive me crazy. It looks like madness - every blogger puts his own tags. The post about how the girl bought herself a new iPad, ran into the store on the way, bought clothes, sat with friends in a cafe, came home, fed the dog and watched the old film - it will be marked differently on different blogs. For example, like this: ipad2, Versace, starbucks, pretty small dog, retromovie or like this: tablet, amazing fashion, coffee, dog like Paris Hilton dog, b / w film. Personally, if I see this, I will decide that this is probably a black and white film about a girl working at a roadside cafe who dreams of a Versace and feeds a dog from the street. This is a serious bug. For this reason, a huge number of articles I need pass by me just because they are marked differently. But that is not all. When I still see the “others“ tag my head finally explodes. Um, guys, do you really think that all people on earth are telepaths? I do not know what they put into this section, but I want to know about it already in advance, I will not click into the unknown. Tags are terminally ill, this is an absolutely accurate diagnosis. Therefore, we need something different, more efficient and simple.
Axiom one: each interprets tags in different ways.
Let's imagine a library in which every reader puts the book not in its place, but where it wanted, because “they have the same roots”, “they are the same height” or “finally a bright orange spot appeared among this gray mass and I not so sad. " Does this seem insane to you? But this is exactly the case with tags. In the end, when everything in the library is completely transformed into chaos, at best tired visitors will hopelessly agree with each other about how to put books, and at worst, they will become overgrown with wool and climb the trees again. Tags were not originally adapted to the network. How many attempts do you need to look at a tag cloud, to guess exactly how many posts in this cloud have two or more tags? Which tags were used for each post? We can understand only what we see - how many tags are there on a blog, and the ratio of different posts. In fact, this is not exactly what we need. We do not need tags, but information that is designated with the help of tags.
Axiom Two: tags are not as informative as we need.
What about the fact that by clicking on the star tag you get posts about celebrities, restaurants, yachts, hotels and geometric shapes, and much more, but not about the stars in the universe that you wanted to find? How to get the traffic service reports about your work when you wanted to know how to increase visits to your site? This is a big problem of the modern Internet, isn't it? A huge number of words has several meanings, and since an AI that is able to understand your specific request does not yet exist, you will continue to receive something that you absolutely do not need.
Axiom Three: the same tag can mean completely different things.
Car, tutu, french fries, nanotechnology. I grab a sedative when I see this. Seriously, I do not understand what this post is about, looking at these tags. Do you think it would be easier for us if we had the opportunity to learn about the proportions in the text, for example: a car (28%), a ballet tutu (32%), french fries (14%), nanotechnology (26%)? I do not think so, firstly, I see only the proportions indicated by tags in the text, and not the information I need in a post.
Secondly, I again have to start interpreting the tags and my guess will be very different from what is in the post. The probability that I guess the first time is very small. And the longer I stand in front of such tags and think about this post, the stronger my desire to leave this page becomes.
Fourth axiom: tags do not give an accurate description of the post.
How did it happen that the tags that were not adapted for the network still got into it outside and turned into a huge octopus, which wrapped its tentacles with the posts we needed so that we could not notice them? Why were the people who started doing this the first to ignore these huge flaws? After all, if you are settling into a library, you will not rearrange books the way you want, but immediately ask - where should each book stand and why is it here and not on the other shelf? What should we do with those who did not think about it? Get him to put a video on youtube on which he apologizes that the world has been in an information impasse for many years or burned at the stake to the cries of "witch, burn her"?
In fact, they are not to blame. They used the direct transfer technique - if it works in the real world, it can work in the network. And if each large company has its own tags for designating its documents, then let everyone put their own on the network. Alas, it was immediately clear that this would not work. Let us recall that some of the first steam locomotives already standing on the rails were moving at the expense of metal legs, because the inventor found it logical to repeat the mechanics of the horse's movement. Let us recall that in the first lethal apparatus the shape of the wings repeated the wings of birds. The metal legs of the tags struggle against the ground with all their might, but they cannot move the train with information, and the flapping of the wings on the sides is more like a mating dance of a male, attracting females than free flight. Let's be honest, the tags don't work. We need another system. Which can work as we need.
Fifth Axiom: tags are just eye candy, not a tool for work.
Decision.
Personally, I see two ways out. I must say that this does not apply to the standardization of tags. It will not make life easier for us - we will continue to think about tags like “car, ballet tutu, french fries, nanotechnologies” standing in the same row - because such combinations do not fit into a logical chain. And we will still beat against a wall of words with multiple meanings like “star.”
The first output is the creation of highly specialized systems. For example, for photos, not search by tags, but search by objects or by forms is much better. We voice with a voice (or enter manually) all the objects we need in the picture and place them in the right places - this is how we formulate a specific query and get a specific or very similar result. The second option - as usual, lies on the surface under the very nose (the end of the first part).
Axiom one: each interprets tags in different ways.
Let's imagine a library in which every reader puts the book not in its place, but where it wanted, because “they have the same roots”, “they are the same height” or “finally a bright orange spot appeared among this gray mass and I not so sad. " Does this seem insane to you? But this is exactly the case with tags. In the end, when everything in the library is completely transformed into chaos, at best tired visitors will hopelessly agree with each other about how to put books, and at worst, they will become overgrown with wool and climb the trees again. Tags were not originally adapted to the network. How many attempts do you need to look at a tag cloud, to guess exactly how many posts in this cloud have two or more tags? Which tags were used for each post? We can understand only what we see - how many tags are there on a blog, and the ratio of different posts. In fact, this is not exactly what we need. We do not need tags, but information that is designated with the help of tags.
Axiom Two: tags are not as informative as we need.
What about the fact that by clicking on the star tag you get posts about celebrities, restaurants, yachts, hotels and geometric shapes, and much more, but not about the stars in the universe that you wanted to find? How to get the traffic service reports about your work when you wanted to know how to increase visits to your site? This is a big problem of the modern Internet, isn't it? A huge number of words has several meanings, and since an AI that is able to understand your specific request does not yet exist, you will continue to receive something that you absolutely do not need.
Axiom Three: the same tag can mean completely different things.
Car, tutu, french fries, nanotechnology. I grab a sedative when I see this. Seriously, I do not understand what this post is about, looking at these tags. Do you think it would be easier for us if we had the opportunity to learn about the proportions in the text, for example: a car (28%), a ballet tutu (32%), french fries (14%), nanotechnology (26%)? I do not think so, firstly, I see only the proportions indicated by tags in the text, and not the information I need in a post.
Secondly, I again have to start interpreting the tags and my guess will be very different from what is in the post. The probability that I guess the first time is very small. And the longer I stand in front of such tags and think about this post, the stronger my desire to leave this page becomes.
Fourth axiom: tags do not give an accurate description of the post.
How did it happen that the tags that were not adapted for the network still got into it outside and turned into a huge octopus, which wrapped its tentacles with the posts we needed so that we could not notice them? Why were the people who started doing this the first to ignore these huge flaws? After all, if you are settling into a library, you will not rearrange books the way you want, but immediately ask - where should each book stand and why is it here and not on the other shelf? What should we do with those who did not think about it? Get him to put a video on youtube on which he apologizes that the world has been in an information impasse for many years or burned at the stake to the cries of "witch, burn her"?
In fact, they are not to blame. They used the direct transfer technique - if it works in the real world, it can work in the network. And if each large company has its own tags for designating its documents, then let everyone put their own on the network. Alas, it was immediately clear that this would not work. Let us recall that some of the first steam locomotives already standing on the rails were moving at the expense of metal legs, because the inventor found it logical to repeat the mechanics of the horse's movement. Let us recall that in the first lethal apparatus the shape of the wings repeated the wings of birds. The metal legs of the tags struggle against the ground with all their might, but they cannot move the train with information, and the flapping of the wings on the sides is more like a mating dance of a male, attracting females than free flight. Let's be honest, the tags don't work. We need another system. Which can work as we need.
Fifth Axiom: tags are just eye candy, not a tool for work.
Decision.
Personally, I see two ways out. I must say that this does not apply to the standardization of tags. It will not make life easier for us - we will continue to think about tags like “car, ballet tutu, french fries, nanotechnologies” standing in the same row - because such combinations do not fit into a logical chain. And we will still beat against a wall of words with multiple meanings like “star.”
The first output is the creation of highly specialized systems. For example, for photos, not search by tags, but search by objects or by forms is much better. We voice with a voice (or enter manually) all the objects we need in the picture and place them in the right places - this is how we formulate a specific query and get a specific or very similar result. The second option - as usual, lies on the surface under the very nose (the end of the first part).
')
Source: https://habr.com/ru/post/123208/
All Articles