|
I'm in the early development (and design) phase of my first big project. Part of this project is planned to utilize the Google Maps API in at least some capacity to at the very least map out business locations. Their API also allows us to pull business "white page" info from Google Places as well which I could use and frankly, had been considering doing.
My question is, when a professional developer is designing and developing a web app, are there any general guidelines when it comes to using APIs and their provided data? What I mean by this is, obviously, when my application uses an API, it has now created a dependency on the organization whose API it uses (in this case Google). Therefore, if, for some reason, Google went down (not likely) or changed their service/model (possible),got hacked/whatever else, my application could break.
On the other hand, if I make less API calls and store more data in my database, that will be more costly on my hardware to be storing redundant data that Google already has out there. Also, chances are, Google is going to have a lot more data that my application will and probably be updated more frequently.
Another consideration is the API costs... I'm not sure whether it would be cheaper to just store business data myself or constantly be sending thousands of requests to Google's servers and having to pay them.
This question is a more general architectural/design question, however I used the Google example to better explain what topics I'm trying to ask about here.
My gut feeling, without any formal education in this area, would be to have my application and database depend on external services/APIs as little as possible for data integrity reasons. However, as I said, I am really taking a shot in the dark here. Even if you don't have a concrete answer for me, could you point me to any references or books that would address my questions? Thank you.
|
|
|
|
|
Graceful degradation.
Functions that don't depend on the "Google API" continue.
When incorporating 3rd part API's, I access them all via my own (wrapper) DLL / Interface; e.g. USPS; UPS; FedEx .... different wrappers; same Interface.
Switched credit card / payment processors by creating a new "wrapper" DLL that had the same interface for the new vendor.
You can also create test / fallback API's in this way.
There are other Search API's out there; just a different "wrapper".
|
|
|
|
|
TheOnlyRealTodd wrote: Therefore, if, for some reason, Google went down (not likely) or changed their service/model (possible),got hacked/whatever else, my application could break.
Err...speaking from actual experience google can refuse to service a request. And of course there can be other communication failures as well. So your calls to google must provide a way to track problems. I suggest logging.
Any time you make a call from one box to another, even if your own box, you need to
1. Expect errors
2. Track errors and as much information as provided about the error.
3. Provide enough information about what you were doing so you can start to determine why it failed.
Also it can also help to log even successes. The fact that the call returned with something that looks like a success doesn't actually mean it is. And such logs can help to determine that the call was made in the first place (no point in blaming google if the call never went to google.)
TheOnlyRealTodd wrote: and store more data in my database, that will be more costly on my hardware to be storing redundant data
Because it will represent terabytes of data? Have you sized your business needs?
Not saying you should do that by the way but when making decisions about performance issues one must have a starting point that is reasonable and realistic based on business (not development) needs.
TheOnlyRealTodd wrote: or constantly be sending thousands of requests to Google's servers and having to pay them.
If one user request generates thousands of calls to google then your are going to have a performance problem which is more important. So again sizing might be relevant.
|
|
|
|
|
Basically someone posted a comment with coding and it causes whoever opens the video link gets redirected to another website (spam) seconds after opening it.
I am unable to remove this comment because it redirects too quickly and I cannot cancel out the loading in any basic ways you think of.
I was able to quickly get a snapshot of the comment responsible for this issue: Http://imgur.com/a/OK0E1
Any ideas to remove this comment by canceling or killing the redirection script or code. I was hoping someone could look at the screenshot and tell me something about that code or script this spammer posted.
|
|
|
|
|
Yes, there are many people who are able to remove it. But unless you give us a link to the message there is not much we can do. And why did you report this under Design and Architecture, and not Spam and Abuse Watch Discussion Boards[^]?
|
|
|
|
|
A quick peek at OP's imgur link says it's not on CP. So our Spam forum isn't appropriate. Probably somewhere in Web Design would be better.
Cheers,
Peter
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
|
Thanks, I did take a look but it made no real sense.
|
|
|
|
|
You need to update your site's code to HTML-encode the comments. Until you do that, your site is vulnerable to persisted cross-site scripting.
Cross-site Scripting (XSS) - OWASP[^]
There are much more dangerous things a malicious user can post. For example, they could steal your cookies, and impersonate you on the site. Or they could download malware onto your users' computers, which would appear to come from your site.
Once you've fixed the vulnerability, you will be able to open the page without the script from that comment executing.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Hi
I would like to develop a direct debit system. I would like to know about how to proceed and which language to use.
Any ideas??
shaheen
|
|
|
|
|
..best idea would be to start talking with a bank.
Bastard Programmer from Hell 
If you can't read my code, try converting it here[^]
|
|
|
|
|
You should start by gathering requirements. Without those, any other consideration is meaningless. It doesn't matter what technology you use if you don't know what you're going to build.
This space for rent
|
|
|
|
|
True, without knowing data and what it should look like in the future it is impossible to predict anything.
|
|
|
|
|
Hi guys,
The app needs to have drag and drop function to build tool with workflow like visio for business users. Easy access for members via web or mobile devices. Internet may not available but mobile can sync data when internet is available.
setting up new application with the following architecture, so need your advice.
html5+JS on frontend on web, and swift on mobile.
Java - MVC webservices provide for both web and mobile front end.
bigdata on the back end.
you know any code generation should be used for Java?
Thanks,
AppNewbie
|
|
|
|
|
If you are new to application development, as your profile name suggests, then I think you are way over your head here, and should start off with something less complex. Just a thought...
|
|
|
|
|
|
Hello everybody.
I have WinForms app with WCF services selfhosted inside.
This app must redirect to some kind of ASP.NET app hosted on server. The user will fill a form and then will redirect again to the WinForms app. What technology to use and how the comunication between winforms app and server on response to bring back focus on winforms opened form ?
What are best practices?
Thank you in advance.
|
|
|
|
|
Windows applications don't have a concept of "redirecting" to a website, and websites cannot launch and control Windows applications.
Depending on what you're trying to do, you might be able to use a WebBrowser control[^] in your Windows application to open the website. You could set the ObjectForScripting property[^] to an object in your application, which would allow javascript on your site to call back into the application via the window.external object.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Another option would be to define fields in your WinForm and POST those directly from your application with HttpClient, rather than trying to move between browser and native app contexts.
"There are three kinds of lies: lies, damned lies and statistics."
- Benjamin Disraeli
|
|
|
|
|
Member 11891291 wrote: I have WinForms app with WCF services selfhosted inside.
I'm a little confused with this one. You have WCF services hosted in your WinForms app? Please explain further.
How about using a database that both your web app and winforms app communicate with. That way both are sharing the data that needs to be common in scope. Your WCF service(s) will facilitate this "sharing", as the same WCF service can be used with both the winform app and the web app.
Just a thought...
|
|
|
|
|
Hello,
Can you put a flat section next to a two point perspective section? In one part of my building it is clearer to show the ramps with the perspective angle. Is it acceptable to line up a flat section with a perspective one or do I keep them both flat?
Thank you
|
|
|
|
|
Sorry, I think you have come to the wrong house.
|
|
|
|
|
To clarify.
This forum is for Design & Architecture of software, not for buildings.
|
|
|
|
|
Use a different color / texture.
|
|
|
|
|
We are using UIP in our project.This an obsolete layer siting inside our application for controlling Session & Navigation management.applicaiton session level object are mainly store in UIP State and Navigation also done using it instead of intrinsic Session Object and Response object although below the layer those are only used. We want to migrate to new technology in phases. Can you please suggest the best approach and technology to replace UIP
|
|
|
|
|
It might help if you told us what "UIP" is.
Google thinks it's either "Usual interstitial pneumonia" or "United International Pictures", but neither of those seem to match what you're describing.
Wikipedia also lists "an open source implementation of the TCP/IP network protocol stack intended for use with tiny 8- and 16-bit microcontrollers"; but again, that doesn't seem to match your question.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
