TRIZ and Captcha
Let's try to solve the task of creating a reliable and easy-to-read captcha using the algorithm for solving inventive problems of ARIZ-85-B .
The application of the algorithm is not entirely correct, since it is designed for physical tasks, but the general principles can be shown on it.
Technical system:
to prevent automatic registrations.
includes: an information element recognized by a person, but not recognizable by a machine.
')
Technical contradiction 1:
If the information is easily read by a person, then it is easy to read by the machine.
Technical contradiction 2:
If the information is difficult to read by the machine, then the person also has difficulty reading
It is necessary:
Provide a simple reading of information by a person who wants to register on a resource and the impossibility of reading information by a machine.
A product is an element that, according to the conditions of the problem, must be processed (manufactured, relocated, modified, improved, protected from harmful effects, detected, measured, etc.). In our case, this is an information element. (I deliberately do not write "picture" in order not to limit the task to the framework in advance)
A tool is an element that directly interacts with a product. In our case, it is the human sense organ and the data entry system of the machine.
A is a man; B - Captcha; B is a car.
We assume that the main function is easy reading by a person, since our ultimate goal is to facilitate registration. Accordingly, we choose the scheme TP-1.
We assume that the easiest person to read in TP-1 is when there is no interference with the reading, that is, captcha is read as plain text.
You must specify:
conflicting pair;
enhanced conflict wording;
what the input X-element should do to solve the problem (what it should preserve and what should eliminate, improve, provide, etc.).
It is given: captcha without hindrances for reading, easily readable by a person, and a machine trying to register. Captcha does not create problems for a person, but does not prevent the registration of a car.
It is necessary: ​​to find such an X-element which, while retaining the captcha's ability not to interfere with a person, would protect the captcha from being read by the machine.
Note: The X-element does not have to be a new part of the system. The X-element is a certain change in the system, a certain X in general.
In the simplest case, the operational zone is the space within which a conflict arises, indicated in the task model.
In our case, the operational zone is an information element of the captcha.
Operational time in our task is the time for which the registration takes place.
We cannot influence the human eye and the captcha machine, therefore internal system CDFs are limited to the CDF of the product — these are the variable characteristics of the captcha — the method of presenting information, different variations of the information presented to the reader.
External system CDRs are limited to the use of a captcha, that is, the web resource where it is used.
Our captcha can:
a) change itself;
b) allow the transition to the super-system (the brick does not change, but the house changes);
c) allow connection with “nothing”, i.e. with emptiness;
d) allow a change for a while.
X-element, without complicating the system at all and without causing harmful phenomena,
eliminates the possibility of registering a car for OB without interfering with the registration of a person.
Let us strengthen the formulation of RBI-1 with an additional requirement: no new elements should be introduced into the system, CDR must be used.
The captcha information element during the time of registration should be easy to read, so as not to impede the registration of a person and not to be easy to read, to prevent the registration of the machine.
In the information element kapchi should be no interference for perception, to ensure easy human readability and there should be interference for perception, to ensure the impossibility of reading the machine.
Interference for perception in captcha should be a hindrance when registering a car and should not be a hindrance when registering a person.
This is not all possible solutions, but just an example: how to try to solve a problem not by trial and error, but using the formalization of the problem and a systematic approach to the solution.
The application of the algorithm is not entirely correct, since it is designed for physical tasks, but the general principles can be shown on it.
Part 1. Task analysis
1.1 Mini task conditions
Technical system:
to prevent automatic registrations.
includes: an information element recognized by a person, but not recognizable by a machine.
')
Technical contradiction 1:
If the information is easily read by a person, then it is easy to read by the machine.
Technical contradiction 2:
If the information is difficult to read by the machine, then the person also has difficulty reading
It is necessary:
Provide a simple reading of information by a person who wants to register on a resource and the impossibility of reading information by a machine.
1.2 Select and write a conflicting pair of elements: the product and the tool.
A product is an element that, according to the conditions of the problem, must be processed (manufactured, relocated, modified, improved, protected from harmful effects, detected, measured, etc.). In our case, this is an information element. (I deliberately do not write "picture" in order not to limit the task to the framework in advance)
A tool is an element that directly interacts with a product. In our case, it is the human sense organ and the data entry system of the machine.
1.3 Let's make graphic diagrams TP-1 and TP-2
A is a man; B - Captcha; B is a car.
1.4 Let us choose from two conflict patterns (TP-1 and TP-2) the one that ensures the best implementation of the main production process (the main function of the technical system specified in the terms of the task).
We assume that the main function is easy reading by a person, since our ultimate goal is to facilitate registration. Accordingly, we choose the scheme TP-1.
1.5 We will increase the conflict by specifying the limiting state (action) of the elements.
We assume that the easiest person to read in TP-1 is when there is no interference with the reading, that is, captcha is read as plain text.
1.6 We write the formulation of the problem model.
You must specify:
conflicting pair;
enhanced conflict wording;
what the input X-element should do to solve the problem (what it should preserve and what should eliminate, improve, provide, etc.).
It is given: captcha without hindrances for reading, easily readable by a person, and a machine trying to register. Captcha does not create problems for a person, but does not prevent the registration of a car.
It is necessary: ​​to find such an X-element which, while retaining the captcha's ability not to interfere with a person, would protect the captcha from being read by the machine.
Note: The X-element does not have to be a new part of the system. The X-element is a certain change in the system, a certain X in general.
Part 2. Analysis of the task model
2.1 Operational zone (OZ)
In the simplest case, the operational zone is the space within which a conflict arises, indicated in the task model.
In our case, the operational zone is an information element of the captcha.
2.2 Operational time (OV)
Operational time in our task is the time for which the registration takes place.
2.3. Real field resources (CDF)
We cannot influence the human eye and the captcha machine, therefore internal system CDFs are limited to the CDF of the product — these are the variable characteristics of the captcha — the method of presenting information, different variations of the information presented to the reader.
External system CDRs are limited to the use of a captcha, that is, the web resource where it is used.
Our captcha can:
a) change itself;
b) allow the transition to the super-system (the brick does not change, but the house changes);
c) allow connection with “nothing”, i.e. with emptiness;
d) allow a change for a while.
Part 3. Definition of RBI and AF
3.1 The formulation of the ideal end result (ICR-1)
X-element, without complicating the system at all and without causing harmful phenomena,
eliminates the possibility of registering a car for OB without interfering with the registration of a person.
Let us strengthen the formulation of RBI-1 with an additional requirement: no new elements should be introduced into the system, CDR must be used.
3.2 The formulation of the physical contradiction at the macro level
The captcha information element during the time of registration should be easy to read, so as not to impede the registration of a person and not to be easy to read, to prevent the registration of the machine.
3.3 The formulation of the physical contradiction at the micro level
In the information element kapchi should be no interference for perception, to ensure easy human readability and there should be interference for perception, to ensure the impossibility of reading the machine.
3.4 The formulation of the perfect end result for RBI-2
Interference for perception in captcha should be a hindrance when registering a car and should not be a hindrance when registering a person.
The results of the application of the system of standards to the solution of the problem formulated in the form of ICR-2
- Separation of conflicting properties in space.
Example: ask to enter as captcha any information from the site on which the registration is carried out, not directly related to the registration form. Suppose the name of the penultimate topic of the forum is 12/25/1997. - Separation of conflicting properties in time
Example: Show captcha not during registration, but sooner or later. Suppose invites - the information required for registration is issued before the registration itself passes. - System transition 1. The entire system is endowed with the property C, and its parts with the anti-C property
Example: Classic captcha with changing the position and shape of letters. Individual letters are read well by a person, and the whole system of letters is perceived badly by the machine. - System transition 2. Integration of homogeneous or inhomogeneous systems into a super-system
Example: a combination of 2 or more types of captcha. - Phase transition 1. Replacing the phase state of a part of the system or the external environment.
Example: Audio Captcha, Video Captcha - Phase transition 2. “dual” phase state
Example: With a high frequency, the elements of characters in captcha flicker. The screenshot for the machine will not give intelligible data, the person will read it without any problems.
This is not all possible solutions, but just an example: how to try to solve a problem not by trial and error, but using the formalization of the problem and a systematic approach to the solution.
Source: https://habr.com/ru/post/50228/
All Articles