To collaborate on a project using a version control system, the first thing you have to do is create a local copy by obtaining information from the repository. The user can then modify the copy. There are two basic operating schemes so that users can contribute their modifications:

• - exclusively: in this scheme, in order to make a change, it is necessary to communicate to the repository the element that you wish to modify and the system will be responsible for preventing another user from modifying said element. Once the modification is made, it is shared with the rest of the collaborators. If you have finished modifying an element then that element is released so that others can modify it.

• - Collaboratively: in this scheme each user modifies the local copy and when the user decides to share the changes the system automatically tries to combine the various modifications. The main problem is the possible appearance of conflicts that must be resolved manually or the possible inconsistencies that arise when modifying the same file by several uncoordinated people. Furthermore, this semantics is not appropriate for binary files.

Team Foundation Server version control allows you to choose either of the two forms of collaboration.

Podemos clasificar los sistemas de control de versiones atendiendo a la arquitectura utilizada para el almacenamiento del código:.

• - Distribuidos: cada usuario tiene su propio repositorio. Los distintos repositorios pueden intercambiar y mezclar revisiones entre ellos. Es frecuente el uso de un repositorio, que está normalmente disponible, que sirve de punto de sincronización de los distintos repositorios locales. Ejemplos: Git y Mercurial.

• - Centralizados: existe un repositorio centralizado de todo el código, del cual es responsable un único usuario (o conjunto de ellos). Se facilitan las tareas administrativas a cambio de reducir flexibilidad, pues todas las decisiones fuertes (como crear una nueva rama) necesitan la aprobación del responsable. Algunos ejemplos son CVS, Subversion o Team Foundation Server.

Advantages of distributed systems

• - You need to be connected to the network less times to carry out operations. This produces greater autonomy and greater speed.

• - Even if the remote repository goes down, people can continue working.

• - By making the different repositories a local replica of the information from the remote repositories to which they connect, the information is highly replicated and therefore the system has fewer problems recovering if, for example, the machine that has the remote repository burns down. Therefore there is less need for backups. However, backups are still necessary to resolve situations in which certain information has not yet been replicated.

• - It allows maintaining cleaner central repositories in the sense that a user can decide that certain changes made by him in the local repository are not relevant for the rest of the users and therefore does not allow that information to be publicly accessible, either because it contains unstable versions, in the process of being encoded or with personalized labels and for the exclusive use of the user.

• - The remote server requires fewer resources than a centralized server would need since much of the work is done by local repositories.

• - As distributed systems are more recent than centralized systems, and as they have more flexibility by having a local repository and one or more remote repositories, these systems have been designed to make the use of branches (creation, evolution and merge) easy and to be able to take full advantage of their potential. For example, branches can be created in the remote repository to fix bugs or create new features. But branches can also be created in local repositories so that users can test and depending on the results, merge them with the main development or not. The branches give great flexibility in the way you work.

Advantages of centralized systems

• - In distributed systems there is less control when working as a team since there is no centralized version of everything that is being done in the project.

• - In centralized systems, versions are identified by a version number. However, in distributed version control systems there are no version numbers, since each repository would have its own revision numbers depending on the changes. Instead each version has an identifier to which a tag can be associated (tag).

El flujo de trabajo de un sistema de control de versiones indica cómo se relacionan los distintos usuarios para colaborar entre sí en la consecución de los objetivos del proyecto.

En los sistemas de control de versiones centralizados el diseño del sistema restringe la forma en que los distintos usuarios colaboran entre sí. Sin embargo en los sistemas de control de versiones distribuidos hay mucha más flexibilidad en la forma de colaborar ya que cada desarrollador puede tanto contribuir como recibir contribuciones. Hay distintos flujos de trabajo, donde cada uno se adapta mejor a cierto tipo de proyectos. Veamos algunos ejemplo habituales que también se pueden combinar para así adaptarse mejor al proyecto concreto (por ejemplo podría usarse en general un flujo centralizado, y usar un gestor de integraciones para ciertas partes críticas y de especial importancia).[3]​.

Centralized workflow

In Centralized Workflow, each developer is a worker node. On the other hand, there is a central remote repository that works as a synchronization point. All worker nodes operate equally on the central remote repository.

If it is a distributed version control system, each worker node consists of a private local repository.

A disadvantage of this way of working is that if two users clone from a central point, and both make changes; Only the first person to submit their changes will be able to do so cleanly. The second developer must first merge his work with that of the first, before submitting it, to avoid overwriting the first's changes. That is, it is necessary to do a previous step since non-fast-forward changes cannot be uploaded (non-fast-forward changes).

Workflow with an Integration Manager

In the workflow with an Integration-Manager Workflow (in English Integration-Manager Workflow) in which each developer has write access to their own public repository and read access to the public repositories of all other users. On the other hand, there is a canonical repository, the 'official' representative of the project.

To contribute to these projects, each developer creates their own public clone of the canonical repository and submits their changes (made in a private repository) to it. To 'upload' their changes to the canonical repository, each developer has to make a request to the person managing it.

The main advantage of this way of working is that you can continue working, and the person managing the canonical repository will be able to recover your changes at any time. Collaborators do not have to wait for their changes to be incorporated into the project, everyone can work at their own pace.

Workflow with Dictator and Lieutenants

The Dictator and Lieutenants Workflow is an extension of the Integration Manager Workflow. It is used in very large projects, with hundreds of collaborators, such as the Linux kernel.

There are a series of integration managers that are in charge of specific parts of the repository, which are called lieutenants. All lieutenants report to an integration manager; known as the dictator. The dictator integrates all the contributions of the lieutenants by publishing the work in a reference repository from which all collaborators retrieve.

This work system allows the group leader (the dictator) to delegate much of the work to the lieutenants, relegating their work to collecting the fruit of multiple work points.

Las ramas, en un sistema de control de versiones, constituyen una potente herramienta que flexibiliza la forma en la que los colaboradores cooperan en el proyecto (en inglés Branching Workflows). Las ramas son solo una herramienta que es posible utilizar de distintas formas para conseguir distintos objetivos. Veamos algunas posibilidades.[3]​[4]​.

long-haul branches

In some projects there are several branches always open, which indicate varying degrees of content stability. For example, in the 'master' branch it is common to maintain only what is completely stable. Then there are other branches that reveal different degrees of stability. For example, we could have a 'beta' branch (beta version) and another 'alpha' branch (alpha version), which we are working on. When a certain degree of stability greater than the branch in which it is present is reached, a fusion is performed with a branch of greater stability.

Specific branches

Specific branches, also called support branches, are branches that are created occasionally to perform a very specific functionality. For example, adding a new feature (they are called new functionality branches or directly by their English name topic branch or feature branch) or correcting a specific bug (they are called bug fix branches or directly by their English name hotfix branch).

These types of branches allow us to work focusing exclusively on the development of a specific feature and when it is completed it is merged with one of the long-term branches (normally with the one with the lowest level of stability, so that it can be tested in that environment). Merging is only done when you are 'sure' that feature is correctly implemented, rather than merging in the order in which things are developed. This allows, on the one hand, to have a history of the different versions that have been had until the functionality was achieved. On the other hand, they allow the history of long-term branches not to be 'dirty' with different modifications related to a specific functionality.

The use of this type of branches allows more flexibility when testing possible solutions. It only merges with long running branches once we are sure we choose the best solution.

One type of branch of this type that has a special function is the so-called version branches or release branches* (in English release branch). These types of branches are created to support the preparation of a new production version. They allow you to have the content of the version under control and to be able to perform certain maintenance on it (add small improvements and correct errors).

It is common and good practice to use a prefix in the branch name that indicates the type of branch it is. For example you could use ‘feature-’ for new feature branches, ‘hotfix-’ for bug-fixing branches, and ‘release-’ for release branches.

• - Version control programs.

• - Software configuration management.

• - Software.

• -CVS.

• - Git.

• - GitHub.

To collaborate on a project using a version control system, the first thing you have to do is create a local copy by obtaining information from the repository. The user can then modify the copy. There are two basic operating schemes so that users can contribute their modifications:

• - exclusively: in this scheme, in order to make a change, it is necessary to communicate to the repository the element that you wish to modify and the system will be responsible for preventing another user from modifying said element. Once the modification is made, it is shared with the rest of the collaborators. If you have finished modifying an element then that element is released so that others can modify it.

• - Collaboratively: in this scheme each user modifies the local copy and when the user decides to share the changes the system automatically tries to combine the various modifications. The main problem is the possible appearance of conflicts that must be resolved manually or the possible inconsistencies that arise when modifying the same file by several uncoordinated people. Furthermore, this semantics is not appropriate for binary files.

Team Foundation Server version control allows you to choose either of the two forms of collaboration.

Podemos clasificar los sistemas de control de versiones atendiendo a la arquitectura utilizada para el almacenamiento del código:.

• - Distribuidos: cada usuario tiene su propio repositorio. Los distintos repositorios pueden intercambiar y mezclar revisiones entre ellos. Es frecuente el uso de un repositorio, que está normalmente disponible, que sirve de punto de sincronización de los distintos repositorios locales. Ejemplos: Git y Mercurial.

• - Centralizados: existe un repositorio centralizado de todo el código, del cual es responsable un único usuario (o conjunto de ellos). Se facilitan las tareas administrativas a cambio de reducir flexibilidad, pues todas las decisiones fuertes (como crear una nueva rama) necesitan la aprobación del responsable. Algunos ejemplos son CVS, Subversion o Team Foundation Server.

Advantages of distributed systems

• - You need to be connected to the network less times to carry out operations. This produces greater autonomy and greater speed.

• - Even if the remote repository goes down, people can continue working.

• - By making the different repositories a local replica of the information from the remote repositories to which they connect, the information is highly replicated and therefore the system has fewer problems recovering if, for example, the machine that has the remote repository burns down. Therefore there is less need for backups. However, backups are still necessary to resolve situations in which certain information has not yet been replicated.

• - It allows maintaining cleaner central repositories in the sense that a user can decide that certain changes made by him in the local repository are not relevant for the rest of the users and therefore does not allow that information to be publicly accessible, either because it contains unstable versions, in the process of being encoded or with personalized labels and for the exclusive use of the user.

• - The remote server requires fewer resources than a centralized server would need since much of the work is done by local repositories.

• - As distributed systems are more recent than centralized systems, and as they have more flexibility by having a local repository and one or more remote repositories, these systems have been designed to make the use of branches (creation, evolution and merge) easy and to be able to take full advantage of their potential. For example, branches can be created in the remote repository to fix bugs or create new features. But branches can also be created in local repositories so that users can test and depending on the results, merge them with the main development or not. The branches give great flexibility in the way you work.

Advantages of centralized systems

• - In distributed systems there is less control when working as a team since there is no centralized version of everything that is being done in the project.

• - In centralized systems, versions are identified by a version number. However, in distributed version control systems there are no version numbers, since each repository would have its own revision numbers depending on the changes. Instead each version has an identifier to which a tag can be associated (tag).

El flujo de trabajo de un sistema de control de versiones indica cómo se relacionan los distintos usuarios para colaborar entre sí en la consecución de los objetivos del proyecto.

En los sistemas de control de versiones centralizados el diseño del sistema restringe la forma en que los distintos usuarios colaboran entre sí. Sin embargo en los sistemas de control de versiones distribuidos hay mucha más flexibilidad en la forma de colaborar ya que cada desarrollador puede tanto contribuir como recibir contribuciones. Hay distintos flujos de trabajo, donde cada uno se adapta mejor a cierto tipo de proyectos. Veamos algunos ejemplo habituales que también se pueden combinar para así adaptarse mejor al proyecto concreto (por ejemplo podría usarse en general un flujo centralizado, y usar un gestor de integraciones para ciertas partes críticas y de especial importancia).[3]​.

Centralized workflow

In Centralized Workflow, each developer is a worker node. On the other hand, there is a central remote repository that works as a synchronization point. All worker nodes operate equally on the central remote repository.

If it is a distributed version control system, each worker node consists of a private local repository.

A disadvantage of this way of working is that if two users clone from a central point, and both make changes; Only the first person to submit their changes will be able to do so cleanly. The second developer must first merge his work with that of the first, before submitting it, to avoid overwriting the first's changes. That is, it is necessary to do a previous step since non-fast-forward changes cannot be uploaded (non-fast-forward changes).

Workflow with an Integration Manager

In the workflow with an Integration-Manager Workflow (in English Integration-Manager Workflow) in which each developer has write access to their own public repository and read access to the public repositories of all other users. On the other hand, there is a canonical repository, the 'official' representative of the project.

To contribute to these projects, each developer creates their own public clone of the canonical repository and submits their changes (made in a private repository) to it. To 'upload' their changes to the canonical repository, each developer has to make a request to the person managing it.

The main advantage of this way of working is that you can continue working, and the person managing the canonical repository will be able to recover your changes at any time. Collaborators do not have to wait for their changes to be incorporated into the project, everyone can work at their own pace.

Workflow with Dictator and Lieutenants

The Dictator and Lieutenants Workflow is an extension of the Integration Manager Workflow. It is used in very large projects, with hundreds of collaborators, such as the Linux kernel.

There are a series of integration managers that are in charge of specific parts of the repository, which are called lieutenants. All lieutenants report to an integration manager; known as the dictator. The dictator integrates all the contributions of the lieutenants by publishing the work in a reference repository from which all collaborators retrieve.

This work system allows the group leader (the dictator) to delegate much of the work to the lieutenants, relegating their work to collecting the fruit of multiple work points.

Las ramas, en un sistema de control de versiones, constituyen una potente herramienta que flexibiliza la forma en la que los colaboradores cooperan en el proyecto (en inglés Branching Workflows). Las ramas son solo una herramienta que es posible utilizar de distintas formas para conseguir distintos objetivos. Veamos algunas posibilidades.[3]​[4]​.

long-haul branches

In some projects there are several branches always open, which indicate varying degrees of content stability. For example, in the 'master' branch it is common to maintain only what is completely stable. Then there are other branches that reveal different degrees of stability. For example, we could have a 'beta' branch (beta version) and another 'alpha' branch (alpha version), which we are working on. When a certain degree of stability greater than the branch in which it is present is reached, a fusion is performed with a branch of greater stability.

Specific branches

Specific branches, also called support branches, are branches that are created occasionally to perform a very specific functionality. For example, adding a new feature (they are called new functionality branches or directly by their English name topic branch or feature branch) or correcting a specific bug (they are called bug fix branches or directly by their English name hotfix branch).

These types of branches allow us to work focusing exclusively on the development of a specific feature and when it is completed it is merged with one of the long-term branches (normally with the one with the lowest level of stability, so that it can be tested in that environment). Merging is only done when you are 'sure' that feature is correctly implemented, rather than merging in the order in which things are developed. This allows, on the one hand, to have a history of the different versions that have been had until the functionality was achieved. On the other hand, they allow the history of long-term branches not to be 'dirty' with different modifications related to a specific functionality.

The use of this type of branches allows more flexibility when testing possible solutions. It only merges with long running branches once we are sure we choose the best solution.

One type of branch of this type that has a special function is the so-called version branches or release branches* (in English release branch). These types of branches are created to support the preparation of a new production version. They allow you to have the content of the version under control and to be able to perform certain maintenance on it (add small improvements and correct errors).

It is common and good practice to use a prefix in the branch name that indicates the type of branch it is. For example you could use ‘feature-’ for new feature branches, ‘hotfix-’ for bug-fixing branches, and ‘release-’ for release branches.

• - Version control programs.

• - Software configuration management.

• - Software.

• -CVS.

• - Git.

• - GitHub.