Support » Plugin: Social Media Feather | social media sharing » Loading speed of plugin compared to Sociable, part 2

  • Resolved wpfan1000


    Dear Synved,

    Please dont mark my issue as Resolved without my agreement.

    I could not undo your marking of my issue of Resolved so I am going to repost.

    My issue that you resolved/dismissed is at:

    For sake of convenience I will repeat my post here:

    In using some web site loading testers, it has been reported that the Social Media Feather plugin is not that fast.

    To get some real data, I did the following:

    1) I installed the Sociable sharing plugin
    1a) I config’d it to display about 10 buttons plus Like buttons
    2) I already had Social Media Feather plugin installed and placed the widget in the sidebar
    3) I installed P3 – Plugin Performance Profiler plugin and ran it.

    Here are the results of loading times:

    Sociable – 0.0326 seconds
    Social Media Feather – 0.1279 seconds

    Ie Social Media Feather is about 4 times slower than Sociable, even though Sociable loads more buttons and gets Like quantity info.

    Could you please comment on this?

    Your response was:

    “That P3 Plugin is fairly unreliable in terms of “performance” measurements. I’m not sure what criteria it uses but they’re not really significant”

    So, you are not sure how it works / what criteria it uses, but you are willing to mark my issue as Resolved without knowing how these results were generated?

    I also wonder what the devs of P3 plugin would have to say about your comment, as well as the users of the 165,504 downloads, and the users who overall rated the plugin at 4.7 out of 5 stars.

    I have asked them to comment since it seems you nor I understand how their plugin works.

    Then you say:
    “Another user brought this up some time ago and came to the same conclusion.”

    The user has this to say:

    “Well, I didn’t use pingdom cuz it costs money, I found a free alternative – it’s not very good, but assuming it’s at least slightly accurate, the average of 5 tests with your plugin enabled and 5 without actually loaded slightly faster with it enabled.

    So pretty inconclusive results but would support your argument that the P3 plugins results were maybe a bit misleading!”

    So in other words:
    1) He did not use pingdom
    2) He did not test your plugin against Sociable
    3) He states “pretty inclusive results”

    So you have marked my issue as Resolved even though:

    1) You not knowing how P3 works (if you, as you say, do not know the criteria it uses, then you do not know how it works)
    2) The other user NOT comparing your plugin against Socialble
    3) The other user NOT using Pingdom
    4) The other user says “Inconclusive”
    5) And lastly, without giving me the opportunity to do further testing, or as me being the original poster, allowing me to mark my own issue as resolved, when I think it would be.

    PS I do not think this issue is Resolved, even though you marked it as so.

    I do a bit of support in the WordPress forum and I would never mark someome else’s issue as resolved in this way.

    I dont appreciate you marking my issue as resolved and frankly I feel you dismissed my issue. You did this based on not knowing how the P3 plugin works, referring me to an inconclusive post, and without providing any real data, explanation or opportunity to discuss.

Viewing 5 replies - 1 through 5 (of 5 total)
  • Hi wpfan1000,

    I’m the author of P3 and am commenting here at your request.

    I don’t know if this should be resolved or not. I can provide some background on what P3 does/is/measures.

    P3 measures what’s happening while WordPress is rendering a the HTML for a page, for a logged in user, then traces that activity back to individual files, and based on the file paths, categorizes the activity into individual plugins, or the theme, or core run time. When you “auto scan”, you’re actually hitting random pages from the front end, plus some admin pages and then averaging it all out. On top of that P3 is designed to bypass caching mechanisms (on top of being performed as a logged-in user). It gives you a very specific view into your performance.

    The method that P3 uses to measure performance is inside php and not in an extension. This is slower and less accurate than other tools. P3 works this way because it doesn’t require a special extension to use and is generally “accurate enough” for a user to decide of a plugin is fast enough for their site.

    Tools like xdebug, xhprof, and newrelic are harder to use, but will give you more accurate results. They require you having the ability to load extensions on the server, though, and you’ll need to know how to interpret the results since they won’t be broken down based on plugins.

    Tools like pingdom and webpagetest will give you a browser’s view into what’s going on. This is important if the plugin is referencing extra scripts / images / styles, and especially when they’re on another domain.

    P3’s results are also not universal. This plugin may perform very well on one user’s setup and perform very poorly on another user’s setup because of differences in server architecture, hardware, traffic, other plugins, etc.

    Bottom line: P3’s results should be a starting point. Next steps would be performing a manual scan in P3 (see the help tab) for better control over the results. Beyond that, you should look up some of the tools mentioned above or corner some devs at a WordCamp and have them show you how to profile code

    I hope this helps!

    @ Synved:

    I used pingdom to test the loading of my site, and I tested my site with your plugin 3 times, to try and get an average.

    The load time varied between 2 – 3 seconds.

    This could of course be for many reasons, including the internet connection between pingdom and my site, server loading at the time of each test, etc.

    So the variation of load times with the same plugin is too great to try and test load times with another plugin and compare the two.

    Ie Load times with your plugin and Sociable plugin could vary by more than a second due to other factors. I dont think there is a way to tell whether load time differences are due to different plugins or because of internet connections, server loading, etc

    So I do not think pingdom is an effective tool to test and compare load times of plugins.

    I will have to research some other ways to try and compare plugins.

    How do you test the loading of your plugin?

    You have to ensure to select the same “server location” in pingdom when you run your tests.

    If you run 10 tests with 1 plugin and the average is 1.5, then with the other plugin the average is 1.8, that’s a decent indicator.

    If you want to compare running time you could place some PHP in your wp-config before and after the include lines at the bottom but again that’s not super reliable unless you run your site on a private server (and even then, it’s not super reliable if the difference is really small).

    Stumbled across this thread when I was checking for resolution on an unrelated issue with this plugin, and it made me shake my head in disgust, and I felt obligated to post.

    I fail to see why OP is making a huge deal about a load time difference of less than a 100ms. This translates to less than a 10th of a second on the user end. I highly doubt it would have any impact on the user experience.

    Seems a lot like nit-picking to me, and being you are being given FREE support on a plugin, you should be grateful to get any response whatsoever. I’m a web dev, and have used this plugin on dozens of sites, and I have YET to hear 1 complaint about speed.

    I currently have a plugin in the repo myself, and if someone started a thread like you did, I’d insta-lock it after laughing in their face.

    Well I dont see many other simple options other than testing with a Pingdom-like site (which to me is not a good way to go as I explained above) or using P3 (which the developer says is a good starting point but there are better options) and so I will mark this post as resolved anyways.

Viewing 5 replies - 1 through 5 (of 5 total)
  • The topic ‘Loading speed of plugin compared to Sociable, part 2’ is closed to new replies.