I'm building an API and I have a question about how to represent objects.
Imagine we have a system with Articles that have a bunch of properties. Some of these properties are complex, for example the Author of the Article refers to another object. We have an URL to fetch all the articles in the system, and another URL to fetch a particular Article.
My first approach to implement this would be to create two representations of the same object Article, because when you request all the articles, it makes sense that you don't retrieve all the information about the Articles, but for example just the title, the date and the name of the author (instead of the whole Author object), excluding other properties like tags, or the content. The idea beneath this is to try to make the response of all the Articles a little bit lighter.
Now I'm going to the client side, and I decide to implement a SDK for Android, for example. So the first step would be to create the objects to store the information that I retrieve from the API. Now a problem pops up, because I want to define the Article object, but I would need two versions of it and it's not only more difficult to implement, but it's going to be more difficult to use.
So my question is, when defining an API, is it a good practice to have multiple versions of the same object (maybe a light one, and a full one) to save some bandwidth when sending the result of a request but generating a more difficult to use service, or it's not worth it and you should retrieve always the same version of the object, generating heavier responses but making the service easier to use?
I work at a company that deals with Articles as well and we also have a REST API to expose the data.
I think you're on the right track, but I'll even take it one step further. These are the potential three calls for large entities in an API:
Index. For the articles, this would be something like /articles. It just returns a list of article ids. You can add parameters to filter, sort, etc. It's very lightweight and I've found it to be very useful.
Header/Mini/Light version. These are only the crucial fields that you think will meet the widest variety of use cases. For us, we have a lot of use cases where we might want to display the top 5 articles, and in those cases, only title, author and maybe publication date. Those fields belong in a "header" article, or a "light" article. This is especially useful for AJAX calls as you don't want to return the entire article (for us the object is quite large.)
Full version. This is the full article. All the text/paragraphs/image references - everything. It's a heavy call to make, but you will be guaranteed to get whatever is available.
Then it just takes discipline to leave the objects the way they are. Ideally users are able to get the version described in (2) to save time over the wire, but if they have to, they go with (3).
I've considered having a dynamic way to return only fields people are interested in, but it would be a lot of implementation. Basically the idea was to let the user go to /article and then show them a sample JSON result. Then the user could click on the fields they wanted returned and get a token. Then they'd pass the token as a parameter to the API and the API would then know which fields to return.
Creates a dynamic schema. Lots of work and I never got around to it, but you can see that if you want to be creative, you can.
Consider whether your data (for one API client) is changing a lot or not. If it's possible to cache data on the client, that'll improve performance by not contacting the API as much. Otherwise I think it's a good idea to have a light-weight and full-scale object type (or more like two views of the same object type).
In the client you should implement it as one object type (to keep it DRY; Don't Repeat Yourself) with all the properties. When fetching a light-weight object, you only store a few of the properties, the rest being null (or similar “undefined” value for the given property type). It should be possible to determine whether all or only a partial subset of the properties are loaded.
When making API requests in the client on a given model (ie. authors) you should be explicit about whether the light-weight or full-scale object is needed and whether cached data is acceptable. This makes it possible to control the data in the UI layer. For example a list of authors might only need to display a name and a number of articles connected with that author. When displaying the author screen, more properties are needed. Also, if using cached data, you should provide a way for the user to refresh it.
When the app works you can start to implement optimizations like: Don't fetch light-weight data if full-scala data is already known & Don't fetch data at all if a recent cache copy exists. I think the best is to look at the actual use cases and improve performance with the highest value for the user.
Related
For example, we have an entity called ServiceConfig that contains a pointer to a Service and a Professional. If returned without including the fields would look like this:
{
'type': '__Pointer',
'className': 'Service',
'objectId': 'q92he840'
}
At which point they could query again to retrieve that service. However, it is often the case that they need the Service name. In which case it is inefficient to have to query again to get the service every time.
Options:
Automatically return the Service. In which case, we should automatically return the Industry for that Service as well in case they need that... same applies to all. Seems like we're returning data too often here.
Allow them to pass an includes parameter that specifies which entities to include. Format is an array of strings where using a . can allow them to include subclasses. In this case ['Professional', 'Service.Industry'] would work.
Can anyone identify why any one solution would be better than the others? I feel that the last solution is the best, however it does not seem to be common to do to in the APIs I've seen.
This is a good API Design decision to spend your time on before you release an initial version. Both your approaches are valid and it all depends on what you think are the most common ways that clients would use your API.
Here are some points that you could consider:
You might prefer the first approach where you do not give all the data upfront. Sometimes it is about efficiency and at times it is also about security and ensuring that any additional important data is only fetched on as as needed basis and on authorization.
Implementing the 2nd approach is going to take more effort on part of your team to design/code and test out the API. So you might want to consider how much of effort you want to put into release 1.0
Since you have nested data for example, the second approach will serve you well. Several public APIs do that as a matter of fact. For e.g. look at the LinkedIn public API and particularly the facets section, where you can specify the fields or additional information that you would like to return.
Look at some of the client applications that you have written and if you can identify for sure that some data is needed anyways upfront, then it can help in designed the return data.
Eventually monitoring API usage and doing some analysis on the number of calls, methods invoked will give you good inputs on what to do next.
If I had to make a choice and have a little bit more leeway in terms of effort, I would go with the 2nd option, even if it is a simple version at first.
My issue is that I'm dealing with a RESTful API that returns information about objects, and when writing classes to represent them, I'm not sure how best to handle all the possibilities of the status of each variable's availability. From what I can tell, there are 5 possibilities: The information
is available
has not been requested
is currently being requested (asynchronously)
is unavailable
is not applicable
So with these, having an object represent its data with a value or null doesn't cut it. To give a more concrete example, I'm working with an API about the United States Congress, so the problem goes as thus:
I request information about a bill, and it contains a stub about the sponsoring legislator.
I eventually need to request all the information about that legislator. Not all the legislators will have all the information. Those in the House of Representatives won't have a senate class (Senators' six-year terms are staggered so a third expire every two years, the House is entirely re-elected every two years). Some won't have a twitter id, just because they don't have one. And, of course, if I have already requested information, I shouldn't try to request it again.
There's a couple options I see:
I can create a Legislator object and fill it with what information I have, but then I have to have some mechanism of tracking information availability with the getters and setters. This is kind of what I'm doing right now, but it requires a lot of repeated code.
I could create a separate class for abbreviated objects and replace them when I get more with immutable "complete" objects, but then I have to be really careful about replacing all references to them and also go through a bunch of hoops for unavailable, and especially, not applicable information.
So, I'm just wondering what other people's take on this issue is. Are there other (better?) ways of handling this complexity? What are the advantages and drawbacks of different approaches? What should I consider about what I'm trying to do in choosing an approach?
[Note: I'm working in Objective-C, but this isn't necessarily specific to that language.]
If you want to treat those remote resources as objects on the client side, the do yourself a huge favour and forget about the REST buzzword. You will drive yourself crazy. Just accept that you are doing HTTP RPC and move on as you would doing any other RPC project.
However, if you really want to do REST, you need to understand what is meant by the "State Transfer" part of the REST acronym and you need to read about HATEOAS. It is a huge mental shift for building clients, but it does have a bunch of benefits. But maybe you don't need those particular benefits.
What I do know, is if you are trying using a "REST API" to retrieve objects over the wire, you are going to come to the conclusion that REST is a load of crap.
It's an interesting question, but I think you're probably overthinking this a bit.
Firstly, I think you're considering the possible states of information a bit too much; consider the more basic consideration that you either have the information or you don't. WHY you have the information doesn't really matter, except in one case. Let me explain; if the information about a certain bill or legislator or anything is not applicable, you shouldn't be requesting it / needing it. That "state" is irrelevant. Similarly, if the information is in the process of being requested, then it is simply not yet available; the only state you really care about is whether you have the information or if you do not yet have the information.
If you start worrying about further depths of the request process, you risk getting into a deep, endless cycle of managing state; has the information changed between when I got it and now? All you can know about the information is if you've been told what it is. This is fundamental to the REST process; you're getting REPRESENTATION of the underlying data, but there's no mistake about it; the representation is NOT the underlying data, any more than a congressman's name is the congressman himself.
Second, don't worry about information availability. If an object has a subobject, when you query the object, query for the subobject. If you get back data, great. If you get back that the data isn't available, that too is a representation of the subobject's data; it's just a different representation than you were hoping for, but it's equally valid. I'd represent that as an object with a null value; the object exists (was instantiated because it belonged to the parent), but you have no valid data about it (the representation returned was empty due to some reason; lack of availability, server down, data changed; whatever).
Finally, the real key here is that you need to be remembering that a RESTful structure is driven by hypermedia; a request to an object that does not return the full object's data should return an URI for requesting the subobject's data; and so forth. The key here is that those structures aren't static, like your object structure seems to be hoping to treat them; they're dynamic, and it's up to the server to determine the representation (i.e., the interrelationship). Attempting to define that in stone with a concrete object representation ahead of time means that you're dealing with the system in a way that REST was never meant to be dealt with.
Closed. This question needs to be more focused. It is not currently accepting answers.
Closed 7 years ago.
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
A REST API can have parameters in at least two ways:
As part of the URL-path (i.e. /api/resource/parametervalue )
As a query argument (i.e. /api/resource?parameter=value )
What is the best practice here? Are there any general guidelines when to use 1 and when to use 2?
Real world example: Twitter uses query parameters for specifying intervals. (http://api.twitter.com/1/statuses/home_timeline.json?since_id=12345&max_id=54321)
Would it be considered better design to put these parameters in the URL path?
If there are documented best practices, I have not found them yet. However, here are a few guidelines I use when determining where to put parameters in an url:
Optional parameters tend to be easier to put in the query string.
If you want to return a 404 error when the parameter value does not correspond to an existing resource then I would tend towards a path segment parameter. e.g. /customer/232 where 232 is not a valid customer id.
If however you want to return an empty list then when the parameter is not found then I suggest using query string parameters. e.g. /contacts?name=dave
If a parameter affects an entire subtree of your URI space then use a path segment. e.g. a language parameter /en/document/foo.txt versus /document/foo.txt?language=en
I prefer unique identifiers to be in a path segment rather than a query parameter.
The official rules for URIs are found in this RFC spec here. There is also another very useful RFC spec here that defines rules for parameterizing URIs.
Late answer but I'll add some additional insight to what has been shared, namely that there are several types of "parameters" to a request, and you should take this into account.
Locators - E.g. resource identifiers such as IDs or action/view
Filters - E.g. parameters that provide a search for, sorting or narrow down the set of results.
State - E.g. session identification, api keys, whatevs.
Content - E.g. data to be stored.
Now let's look at the different places where these parameters could go.
Request headers & cookies
URL query string ("GET" vars)
URL paths
Body query string/multipart ("POST" vars)
Generally you want State to be set in headers or cookies, depending on what type of state information it is. I think we can all agree on this. Use custom http headers (X-My-Header) if you need to.
Similarly, Content only has one place to belong, which is in the request body, either as query strings or as http multipart and/or JSON content. This is consistent with what you receive from the server when it sends you content. So you shouldn't be rude and do it differently.
Locators such as "id=5" or "action=refresh" or "page=2" would make sense to have as a URL path, such as mysite.com/article/5/page=2 where partly you know what each part is supposed to mean (the basics such as article and 5 obviously mean get me the data of type article with id 5) and additional parameters are specified as part of the URI. They can be in the form of page=2, or page/2 if you know that after a certain point in the URI the "folders" are paired key-values.
Filters always go in the query string, because while they are a part of finding the right data, they are only there to return a subset or modification of what the Locators return alone. The search in mysite.com/article/?query=Obama (subset) is a filter, and so is /article/5?order=backwards (modification). Think about what it does, not just what it's called!
If "view" determines output format, then it is a filter (mysite.com/article/5?view=pdf) because it returns a modification of the found resource rather than homing in on which resource we want. If it instead decides which specific part of the article we get to see (mysite.com/article/5/view=summary) then it is a locator.
Remember, narrowing down a set of resources is filtering. Locating something specific within a resource is locating... duh. Subset filtering may return any number of results (even 0). Locating will always find that specific instance of something (if it exists). Modification filtering will return the same data as the locator, except modified (if such a modification is allowed).
Hope this helped give people some eureka moments if they've been lost about where to put stuff!
It depends on a design. There are no rules for URIs at REST over HTTP (main thing is that they are unique). Often it comes to the matter of taste and intuition...
I take following approach:
url path-element: The resource and its path-element forms a directory traversal and a subresource (e.g. /items/{id} , /users/items). When unsure ask your colleagues, if they think that traversal and they think in "another directory" most likely path-element is the right choice
url parameter: when there is no traversal really (search resources with multiple query parameters are a very nice example for that)
IMO the parameters should be better as query arguments. The url is used to identify the resource, while the added query parameters to specify which part of the resource you want, any state the resource should have, etc.
As per the REST Implementation,
1) Path variables are used for the direct action on the resources, like a contact or a song
ex..
GET etc /api/resource/{songid} or
GET etc /api/resource/{contactid} will return respective data.
2) Query perms/argument are used for the in-direct resources like metadata of a song
ex..,
GET /api/resource/{songid}?metadata=genres it will return the genres data for that particular song.
"Pack" and POST your data against the "context" that universe-resource-locator provides, which means #1 for the sake of the locator.
Mind the limitations with #2. I prefer POSTs to #1.
note: limitations are discussed for
POST in Is there a max size for POST parameter content?
GET in Is there a limit to the length of a GET request? and Max size of URL parameters in _GET
p.s. these limits are based on the client capabilities (browser) and server(configuration).
According to the URI standard the path is for hierarchical parameters and the query is for non-hierarchical parameters. Ofc. it can be very subjective what is hierarchical for you.
In situations where multiple URIs are assigned to the same resource I like to put the parameters - necessary for identification - into the path and the parameters - necessary to build the representation - into the query. (For me this way it is easier to route.)
For example:
/users/123 and /users/123?fields="name, age"
/users and /users?name="John"&age=30
For map reduce I like to use the following approaches:
/users?name="John"&age=30
/users/name:John/age:30
So it is really up to you (and your server side router) how you construct your URIs.
note: Just to mention these parameters are query parameters. So what you are really doing is defining a simple query language. By complex queries (which contain operators like and, or, greater than, etc.) I suggest you to use an already existing query language. The capabilities of URI templates are very limited...
As a programmer often on the client-end, I prefer the query argument. Also, for me, it separates the URL path from the parameters, adds to clarity, and offers more extensibility. It also allows me to have separate logic between the URL/URI building and the parameter builder.
I do like what manuel aldana said about the other option if there's some sort of tree involved. I can see user-specific parts being treed off like that.
There are no hard and fast rules, but the rule of thumb from a purely conceptual standpoint that I like to use can briefly be summed up like this: a URI path (by definition) represents a resource and query parameters are essentially modifiers on that resource. So far that likely doesn't help... With a REST API you have the major methods of acting upon a single resource using GET, PUT, and DELETE . Therefore whether something should be represented in the path or as a parameter can be reduced to whether those methods make sense for the representation in question. Would you reasonably PUT something at that path and would it be semantically sound to do so? You could of course PUT something just about anywhere and bend the back-end to handle it, but you should be PUTing what amounts to a representation of the actual resource and not some needlessly contextualized version of it. For collections the same can be done with POST. If you wanted to add to a particular collection what would be a URL that makes sense to POST to.
This still leaves some gray areas as some paths could point to what amount to children of parent resources which is somewhat discretionary and dependent on their use. The one hard line that this draws is that any type of transitive representation should be done using a query parameter, since it would not have an underlying resource.
In response to the real world example given in the original question (Twitter's API), the parameters represent a transitive query that filters on the state of the resources (rather than a hierarchy). In that particular example it would be entirely unreasonable to add to the collection represented by those constraints, and further that query would not be able to be represented as a path that would make any sense in the terms of an object graph.
The adoption of this type of resource oriented perspective can easily map directly to the object graph of your domain model and drive the logic of your API to the point where everything works very cleanly and in a fairly self-documenting way once it snaps into clarity. The concept can also be made clearer by stepping away from systems that use traditional URL routing mapped on to a normally ill-fitting data model (i.e. an RDBMS). Apache Sling would certainly be a good place to start. The concept of object traversal dispatch in a system like Zope also provides a clearer analog.
Here is my opinion.
Query params are used as meta data to a request. They act as filter or modifier to an existing resource call.
Example:
/calendar/2014-08-08/events
should give calendar events for that day.
If you want events for a specific category
/calendar/2014-08-08/events?category=appointments
or if you need events of longer than 30 mins
/calendar/2014-08-08/events?duration=30
A litmus test would be to check if the request can still be served without an query params.
I generally tend towards #2, As a query argument (i.e. /api/resource?parameter=value ).
A third option is to actually post the parameter=value in the body.
This is because it works better for multi parameter resources and is more extendable for future use.
No matter which one you pick, make sure you only pick one, don't mix and match. That leads towards a confusing API.
One "dimension" of this topic has been left out yet it's very important: there are times when the "best practices" have to come into terms with the plaform we are implementing or augmenting with REST capabilities.
Practical example:
Many web applications nowadays implement the MVC (Model, View, Controller) architecture. They assume a certain standard path is provided, even more so when those web applications come with an "Enable SEO URLs" option.
Just to mention a fairly famous web application: an OpenCart e-commerce shop.
When the admin enables the "SEO URLs" it expects said URLs to come in a quite standard MVC format like:
http://www.domain.tld/special-offers/list-all?limit=25
Where
special-offers is the MVC controller that shall process the URL (showing the special-offers page)
list-all is the controller's action or function name to call. (*)
limit=25 is an option, stating that 25 items will be shown per page.
(*) list-all is a fictious function name I used for clarity. In reality, OpenCart and most MVC frameworks have a default, implied (and usually omitted in the URL) index function that gets called when the user wants a default action to be performed. So the real world URL would be:
http://www.domain.tld/special-offers?limit=25
With a now fairly standard application or frameworkd structure similar to the above, you'll often get a web server that is optimized for it, that rewrites URLs for it (the true "non SEOed URL" would be: http://www.domain.tld/index.php?route=special-offers/list-all&limit=25).
Therefore you, as developer, are faced into dealing with the existing infrastructure and adapt your "best practices", unless you are the system admin, know exactly how to tweak an Apache / NGinx rewrite configuration (the latter can be nasty!) and so on.
So, your REST API would often be much better following the referring web application's standards, both for consistency with it and ease / speed (and thus budget saving).
To get back to the practical example above, a consistent REST API would be something with URLs like:
http://www.domain.tld/api/special-offers-list?from=15&limit=25
or (non SEO URLs)
http://www.domain.tld/index.php?route=api/special-offers-list?from=15&limit=25
with a mix of "paths formed" arguments and "query formed" arguments.
I see a lot of REST APIs that don't handle parameters well. One example that comes up often is when the URI includes personally identifiable information.
http://software.danielwatrous.com/design-principles-for-rest-apis/
I think a corollary question is when a parameter shouldn't be a parameter at all, but should instead be moved to the HEADER or BODY of the request.
It's a very interesting question.
You can use both of them, there's not any strict rule about this subject, but using URI path variables has some advantages:
Cache:
Most of the web cache services on the internet don't cache GET request when they contains query parameters.
They do that because there are a lot of RPC systems using GET requests to change data in the server (fail!! Get must be a safe method)
But if you use path variables, all of this services can cache your GET requests.
Hierarchy:
The path variables can represent hierarchy:
/City/Street/Place
It gives the user more information about the structure of the data.
But if your data doesn't have any hierarchy relation you can still use Path variables, using comma or semi-colon:
/City/longitude,latitude
As a rule, use comma when the ordering of the parameters matter, use semi-colon when the ordering doesn't matter:
/IconGenerator/red;blue;green
Apart of those reasons, there are some cases when it's very common to use query string variables:
When you need the browser to automatically put HTML form variables into the URI
When you are dealing with algorithm. For example the google engine use query strings:
http:// www.google.com/search?q=rest
To sum up, there's not any strong reason to use one of this methods but whenever you can, use URI variables.
I'm currently (I try to) designing a RESTful API for a social network. But I'm not sure if my current approach does still accord to the RESTful principles. I'd be glad if some brighter heads could give me some tips.
Suppose the following URI represents the name field of a user account:
people/{UserID}/profile/fields/name
But there are almost hundred possible fields. So I want the client to create its own field views or use predefined ones. Let's suppose that the following URI represents a predefined field view that includes the fields "name", "age", "gender":
utils/views/field-views/myFieldView
And because field views are kind of higher logic I don't want to mix support for field views into the "people/{UserID}/profile/fields" resource. Instead I want to do the following:
utils/views/field-views/myFieldView/{UserID}
Another example
Suppose we want to perform some quantity operations (hope that this is the right name for it in English). We have the following URIs whereas each of them points to a list of persons -- the friends of them:
GET people/exampleUID-1/relationships/friends
GET people/exampleUID-2/relationships/friends
And now we want to find out which of their friends are also friends of mine. So we do this:
GET people/myUID/relationships/intersections/{Value-1};{Value-2}
Whereas "{Value-1/2}" are the url encoded values of "people/exampleUID-1/friends" and "people/exampleUID-2/friends". And then we get back a representation of all people which are friends of all three persons.
Though Leonard Richardson & Sam Ruby state in their book "RESTful Web Services" that a RESTful design is somehow like an "extreme object oriented" approach, I think that my approach is object oriented and therefore accords to RESTful principles. Or am I wrong?
When not: Are such "object oriented" approaches generally encouraged when used with care and in order to avoid query-based REST-RPC hybrids?
Thanks for your feedback in advance,
peta
I've never worked with REST, but I'd have assumed that GETting a profile resource at '''/people/{UserId}/profile''' would yield a document, in XML or JSON or something, that includes all the fields. Client-side I'd then ignore the fields I'm not interested in. Isn't that much nicer than having to (a) configure a personalised view on the server or (b) make lots of requests to fetch each field?
Hi peta,
I'm still reading through RESTful Web Services myself, but I'd suggest a slightly different approach than the proposed one.
Regarding the first part of your post:
utils/views/field-views/myFieldView/{UserID}
I don't think that this is RESTful, as utils is not a resource. Defining custom views is OK, however these views should be (imho) a natural part of your API's URI scheme. To incorporate the above into your first URI example, I would propose one of the following examples instead of creating a special view for it:
people/{UserID}/profile/fields/name,age,gender/
people/{UserID}/profile/?fields=name,age,gender
The latter example considers fields as an input value for your algorithm. This might be a better approach than having fields in the URI as it is not a resource itself - it just puts constraints on the existing view of people/{UserID}/profile/. Technically, it's very similar as pagination, where you would limit a view by default and allow clients to browse through resources by using ?page=1, ?page=2 and so on.
Regarding the second part of your post:
This is a more difficult one to crack.
First:
Having intersection in the URI breaks your URI scheme a bit. It's not a resource by itself and also it sits on the same level as friends, whereas it would be more suitable one level below or as an input value for your algorithm, i.e.
GET people/{UserID}/relationships/friends/intersections/{Value-1};{Value-2}
GET people/{UserID}/relationships/friends/?intersections={Value-1};{Value-2}
I'm again personally inclined to the latter, because similarly as in the first case, you are just constraining the existing view of people/{UserID}/relationships/friends/
Secondly, regarding:
Whereas "{Value-1/2}" are the url
encoded values of
"people/exampleUID-1/friends" and
"people/exampleUID-2/friends"
If you meant that {Value-1/2} contain the whole encoded response of the mentioned GET requests, then I would avoid that - I don't think that the RESTful way. Since friends is a resource by itself, you may want to expose it and access it directly, i.e.:
GET friends/{UserID-1};{UserID-2};{UserID-3}
One important thing to note here - I've used ; between user IDs in the previous example, whereas I used , in the fields example above. The reasoning is that both represent a different operator. In the first case we needed OR (,) in order to get all three fields, while in the last example above we had to use AND (;) in order to get an intersection.
Usage of two types of operators can over-complicate the API design, but it should provide more flexibility in the end.
thanks for your clarifying answers. They are exactly what I was asking for. Unfortunately I hadn't the time to read "RESTful Web Services" from cover to cover; but I will catch it up as soon as possible. :-)
Regarding the first part of my post:
You're right. I incline to your first example, and without fields. I think that the I don't need it at all. (At the moment) Why do you suggest the use of OR (,) instead of AND (;)? Intuitively I'd use the AND operator because I want all three of them and not just the first one existing. (Like on page 121 the colorpairs example)
Regarding the second part:
With {Value-1/2} I meant only the url-encoded value of the URIs -- not their response data. :) Here I incline with you second example. Here it should be obvious that under the hood an algorithm is involed when calculating intersecting friends. And beside that I'm probably going to add some further operations to it.
peta
Since there aren no respositories for value objects. How can I load all value objects?
Suppose we are modeling a blog application and we have this classes:
Post (Entity)
Comment (Value object)
Tag (Value object)
PostsRespository (Respository)
I Know that when I save a new post, its tags are saved with it in the same table. But how could I load all tags of all posts. Should PostsRespository have a method to load all tags?
I usually do it, but I want to know others opinions
I'm looking for a better solution for this question and I found this post:
http://gojko.net/2009/09/30/ddd-and-relational-databases-the-value-object-dilemma/
This post explain very well why there is a lot of confusion with value objects and databases.
Here you are phrase which liked me too much:
"Persistence is not an excuse to turn everything to entities."
Gojko Adzic, give us three alternatives to save our value objects.
I am currently working through a similar example. Once you need to uniquely refer to tags they are no long simple value objects and may continue to grow in complexity. I decided to make them their own entities and create a separate repository to retrieve them. In most scenarios they are loaded or saved with the post but when they are required alone the other repository is used.
I hope this helps.
EDIT:
In part thanks to this post I decided to restructure my application slightly. You are right that I probably was incorrectly making tags an entity. I have since changed my application so that tags are just strings and the post repository handles all the storage requirements around tags. For operations that need posts the tags are loaded with them. For any operation that just requires tags or lists of tags the repository has methods for that.
Here is my take in how I might solve this type of problem in the way that I'm currently practicing DDD.
If you are editing something that requires tags to be added and removed from such as a Post then tags may be entities but perhaps they could be value objects and are loaded and saved along with the post either way. I personally tend to favor value objects unless the object needs to be modified but I do realize that there is a difference between entity object modeled as read only "snapshots" and actual value objects that lack identity. The tricky part is that perhaps sometimes what you would normally think of as a key could be part of a value object as long as it is not used as identity in that context and I think tags fall into this category.
If you are editing the tags themselves then it is probably a separate bounded context or at least a separate aggregate in which tags are themselves are the aggregate root and persisted through a repository. Note that the entity class that represents tags in this context doesn't have to be the same entity class for tags used in Post aggregate.
If your listing available tags on the display for read only purposes such as to provide a selection list, then that is probably a list of value objects. These value objects can but don't have to be in the Domain Model since they are mainly about supporting the UI and not about the actual domain.
Please chime in if anybody has any thoughts on why my take on this might be wrong but this is the way I've been doing it.