1. Advertising
    y u no do it?

    Advertising (learn more)

    Advertise virtually anything here, with CPM banner ads, CPM email ads and CPC contextual links. You can target relevant areas of the site and show ads based on geographical location of the user if you wish.

    Starts at just $1 per CPM or $0.10 per CPC.

Why I hate WordPress - it's the update thing

Discussion in 'WordPress' started by Jeffr2014, Feb 22, 2015.

  1. #1
    WordPress is becoming the new Microsoft Windows in terms of security issues, with so many websites running on WP everyone tries to hack it. As a result, site owners have to constantly update WordPress versions, themes, and plugins. It is becoming a maintenance nightmare, plus there is absolutely no guarantee that these components work together - you are at mercy of theme and plugin authors. What's worse, every popular plugin becomes more and more bloated, takes more and more CPU and slows down page loading.

    I just had to upgrade 3 plugins that I use in many of my blogs: Add Meta Data, Crafty Social Buttons, and Ultimate Tag Cloud. Noticed that homepage loading become slower and checked with P3. The result? For each one of these three plugins the upgraded version is using 50% more CPU resources that the previous one. Looks like it's time to ditch these and look for alternatives :(
     
    Jeffr2014, Feb 22, 2015 IP
  2. Revelations-Decoder

    Revelations-Decoder Well-Known Member

    Messages:
    3,028
    Likes Received:
    152
    Best Answers:
    4
    Trophy Points:
    190
    #2
    Add Meta Data is one of my favs, I don't suppose you noted how much extra that has added did you please?
     
    Revelations-Decoder, Feb 22, 2015 IP
  3. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #3
    I did, but the value of some of (or rather most of) this metadata is questionable. At the end of the day the big question remains: with regard to SERP, whether this extra metadata is worth 0.05-0.1 extra seconds in page loading time? Nobody but Google can answer this :)
     
    Last edited: Feb 22, 2015
    Jeffr2014, Feb 22, 2015 IP
  4. billzo

    billzo Well-Known Member

    Messages:
    961
    Likes Received:
    278
    Best Answers:
    15
    Trophy Points:
    113
    #4
    The big question: How many of your users are going to notice 5/100ths of a second longer load time? None. If you think Google is going to rank your site higher or lower because of a 5/100ths or 10/100ths of a second load time difference, you are mistaken. You are aware that load time varies by distance from the server and according to network traffic which varies by time of day? Not to mention by whatever processes may be using CPU on the server at that instant.

    Google is crooked and started that whole load time ranking BS because they are trying to sell their own cloud platform. It's all about them driving business to their cloud platform. Sadly, people will fall for it.

    I just happened to be looking at Wordpress description and keyword meta before I read this post. Based on a cursory glance, it looks quite simple to incorporate into a theme using custom post fields without a plugin. I will have to look at it further. One of my biggest complaints about Wordpress is the lack of description meta.
     
    billzo, Feb 22, 2015 IP
  5. Nigel Lew

    Nigel Lew Notable Member

    Messages:
    4,642
    Likes Received:
    405
    Best Answers:
    21
    Trophy Points:
    295
    #5
    What are you using to glean these metrics?

    Nigel
     
    Nigel Lew, Feb 22, 2015 IP
  6. Revelations-Decoder

    Revelations-Decoder Well-Known Member

    Messages:
    3,028
    Likes Received:
    152
    Best Answers:
    4
    Trophy Points:
    190
    #6
    I think I will go with the 0.05-0.1 loss compared to all the benefits "Add Meta Tags" gives me compared to all the other SEO plugins (which I have tried over the years one by one and chosen Add Meta Tags over all others favs). This time lapse is so negligible for the amount of benefits I get with useful meta data if I set that plugin right it's well worth it IMHO.

    I have many page 1 and #1 SERPS as do clients if mine using it so it doesn't seem to be causing any problems from my experience! @Jeffr2014 can you show any data that shows loss of SERPS like you are on page 2 (or below)of say Google or Bing instead of page 1 because of Add Meta Tags plugin please?

    And as @Nigel Lew mentioned where are you getting this data from please?
     
    Revelations-Decoder, Feb 23, 2015 IP
  7. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #7
    1/10 of a second TTFB multiply by 10 plugins is 1 sec. This is (1) noticeable and (2) affects SERP.

    I am running a script that is using webpagetest REST API to check performance of my sites every 15 minutes and outputs stats by averaging loading times over 24 hours period (i.e. average of 360 samples). For impact of different plugins on a server (CPU and DB queries) I use P3 plugin from GoDaddy (re my statement on 50% increase after upgrade).
     
    Jeffr2014, Feb 23, 2015 IP
  8. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #8
    Not sure why are you so eagerly defending Add Meta Tags plugin, I was just giving it as an a example. Speaking of this plugin, I just got rid of it yesterday, it was quite easy to edit a child theme php files to include all basic metadata that this plugin is providing. I just found this post https://wordpress.org/support/topic/efficiency-of-the-new-version-not-happy, looks like I am not the only one who is not happy with the new version of this plugin. Except that I always had Opengraph, Dublin Core, and schema.org metadata disabled - I don't believe SEO benefits of this metadata compensate for added extra 100 ms in TTFB.

    Finally, the post is more about the general issue with WordPress concept of themes and plugins: maintenance, security and stability. Some of my sites get up to 100 hacking attempts daily. In the past 3 months plugin or theme updates broke my sites twice... It is becoming a real nightmare to properly manage WP sites.
     
    Jeffr2014, Feb 23, 2015 IP
  9. Revelations-Decoder

    Revelations-Decoder Well-Known Member

    Messages:
    3,028
    Likes Received:
    152
    Best Answers:
    4
    Trophy Points:
    190
    #9
    Not sure where you see eagerness in finding such a negligible amount OK but you certainly seem to have a bee in your bonnet flaming WP so I will let you carry on.

    The bits you mention you disable I mostly enable but it's horses for courses and perhaps WP is not the course for you. You could try Joomla or Drupal or some other CMS or perhaps another blogging system if it's the blogging you need from it....

    100 hacking attempts a day is about what my hosts say we are getting per hour across a few sites but what can you do.

    Myself I avoid a lot by not installing WP where or how expected which along with other things helps a bit but these days hacking is part and parcel of being online using a CMS type installation like virus attempts are using Windows or other OS's

    If you have evidence of losing SERPS by using Add Meta Tags or any other SEO plugin (like All in One SEO or WordPress SEO or SEO Ultimate) that gives meta data I would certainly be interested in seeing that evidence but tbh your beliefs and mine differ and never the twain shall meet so not a lot of point discussing SEO plugins if it's WP you are (as you say) bemoaning the pitfalls of.
     
    Revelations-Decoder, Feb 23, 2015 IP
  10. stephan2307

    stephan2307 Well-Known Member

    Messages:
    1,277
    Likes Received:
    33
    Best Answers:
    7
    Trophy Points:
    150
    #10
    Interesting read and here are my thoughts.

    In terms of the hacking issue, choosing a good host can take a lot of worry of you shoulders. If you look at wpengine for example they have their own build in security system and they are blocking them by default. Obviously this comes at a price.

    At the same time it is crucial to look through the code of plugins and themes. That way you can identify problems and work with the developer to solve them. You will be surprised how happy they will be for your help to improve the plugin.

    Also I would propose to write your own plugins. Sure there are some plugins that would be tricky to replicate ie woocommerce. But for plugins that do small tasks, write your own plugins. That way the plugin will always do exactly what you want it to do.
     
    stephan2307, Feb 23, 2015 IP
  11. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #11
    Well, I am not that hardcore :) May be in the next incarnation I will write my own plugins... grow my vegetables, make my clothes, etc :)
     
    Jeffr2014, Feb 23, 2015 IP
    neteater likes this.
  12. Revelations-Decoder

    Revelations-Decoder Well-Known Member

    Messages:
    3,028
    Likes Received:
    152
    Best Answers:
    4
    Trophy Points:
    190
    #12
    You could always not use WordPress which as it is being used on some 40% plus websites is bound to be the most hacked but where is the proofs we have asked for that this 0.05 speed from the likes of Add Meta tags loses SERPS?

    Could you please show examples to prove this as I have many page 1 and #1 SERPS that would suggest not using Add Meta tags would not make my SERPS higher as by using it I can still stay number 1 so not sure if your figures have any meaning unless you can show what would prove this to happen....
     
    Last edited: Feb 24, 2015
    Revelations-Decoder, Feb 24, 2015 IP
  13. neteater

    neteater Well-Known Member

    Messages:
    734
    Likes Received:
    5
    Best Answers:
    0
    Trophy Points:
    105
    #13
    LOL
     
    neteater, Feb 25, 2015 IP
  14. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #14
    Dear @Revelations-Decoder , I really don't understand what proves you want from me or why do you take this thread off on tangent. I am sure you are well aware that page loading time and especially TTFB affect SERP (if not then Google it - you'll find many articles with research and supporting data). It is not 0.05-0.1sec per plugin that concerns me, it's cumulative delay for 15-20 plugins that are typically installed at a WP site. For some of my WP sites plugins are responsible for 80% page load delay with 50+ DB queries done by plugins alone.

    Yet, the thread is about STABILITY, SECURITY, and MAINTENANCE of WordPress sites - read the first message.
     
    Jeffr2014, Feb 25, 2015 IP
  15. George Notaras

    George Notaras Greenhorn

    Messages:
    6
    Likes Received:
    1
    Best Answers:
    0
    Trophy Points:
    13
    #15
    Hey all,

    Add-Meta-Tags developer here. First of all, thanks for using this plugin and also for the kind words about it!

    I'd like to post the following remarks about the claims of the OP.

    The claim that enabling Add-Meta-Tags (AMT) adds about 50-100ms of delay in page load time is completely unrealistic and cannot be reproduced in my tests. I tested a rather old version of the plugin, v2.4.3 (1 year old), and the latest release, v2.6.11, using the aforementioned P3 plugin. The results are as follows: v2.4.3: about 14-15ms -- v2.6.11: about 15-17ms. The enhancements of the plugin during the last 12 months completely justify this small increase in processing time. Even the OP's screenshots posted on the aforementioned topic on the WordPress Support forums prove that despite the claims, the delay he is getting is actually around 16-23ms.

    Another important issue is that using the webpagetest service in order to measure the generation time of a dynamic page is a totally inefficient method. There are too many external network-related factors that affect the final result that render the test inaccurate. Such results are not good even as rough estimates. The page load time as measured by the webpagetest service and the page load time measured by P3 plugin are two entirely different things.

    Also, other factors such as oversold shared resources (both in shared-hosting and vps environments) could have a negative impact in these results.

    Here is what I would like to see from OP: 1) P3 results with and without Add-Meta-Tags 2) links to webpagetest org results with and without the plugin.

    If you are unable to provide at least such evidence to back your claims, please stop spreading false information about the plugin.

    Thank you in advance.
     
    Last edited: Mar 1, 2015
    George Notaras, Mar 1, 2015 IP
    Revelations-Decoder likes this.
  16. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #16
    I absolutely hate getting into argument here at DP, especially with a person who volunteers his time to write free plugins... George, I understand your feelings, but I don't like the offensive tone of your complaints (i.e. "please stop spreading false information"). I agree with you that "The page load time as measured by the webpagetest service and the page load time measured by P3 plugin are two entirely different things."
    Yet, both of these metrics are applicable ONLY to a specific site (or rather specific hosting plan on a particular host). As an example, for two sites with the same set of plugins P3 will show 10x difference in page loading times (and timing for a particular plugin) if one is hosed on a dedicated server and another on slow shared hosting. The ONLY reliable data that could be derived from P3 is the relative effect of different plugins on page load time and the total # of db queries. Moreover, even for any specific site you CANNOT assume that the numbers shown by P3 plugin represent TOTAL portion of page loading time by this plugin. We did many tests using webpagetest.org and P3 with activating and deactivating some plugins and calculating their effect (this is how we decide which plugins to use our blog network). The numbers (delta between time average with plugin activated and deactivated) are typically 2-5 times greater than "time" for this plugin shown by P3.
    Anyway, the lessons to take home are:
    - P3 numbers are not transferable between sites, so comparing numbers (i.e. claiming that in my tests for my site with P3 I got 15ms for this plugin and in your tests for your site you got 30ms) is nonsense.
    - For any given site the numbers shown by P3 are "ideal case", the "real life effect" of a plugin on page load time (as could be measured by WPT.org) is typically a few times greater.


    As for the post at WordPress Support forums, I just checked and, apparently, it is by one of our employees - I'll leave it to him to deal with the numbers.

    Once again, the point of original post had nothing to do with AMT plugin, and I regret that we went off tangent here to discuss it. I honestly don't have anything against AMT, it is indeed one of the better plugins that support metadata.
     
    Last edited: Mar 2, 2015
    Jeffr2014, Mar 2, 2015 IP
  17. George Notaras

    George Notaras Greenhorn

    Messages:
    6
    Likes Received:
    1
    Best Answers:
    0
    Trophy Points:
    13
    #17
    What I've written above still stands. Please, stop spreading false information about the plugin, unless you back your claims with solid proof or detailed instructions about how to reproduce the issue. A delay of 50-100ms would be easily reproducible.

    I've asked for the version number of the plugin with which you did not experience the amount of delay you claim you have experienced with v2.6.7. Your employee replied that you do not have this piece of info. Now, what do you expect me to do? Test all releases myself?

    Any test has a degree of reliability as long as it is performed on the same server & network and definitely not on a production box.
    I've repeatedly tested 2.4.3 (12 months old release) and 2.6.11 (latest release) on the same non-production box with dedicated resources using the P3 plugin. That is an as reliable as possible method to measure the overhead of using the plugin. I did not find significant differences or differences I could not justify between the two versions.

    You're trying to measure the impact the plugin has on the page load time by measuring your web sites' response time through WPT, which combines the DNS query, the generation of the dynamic page on a live website on a production server and the transfer of the response over the internet. Well, you are doing it wrong! How do you know if the server or the network are not overloaded? The "the real life effect of a plugin" you referred to does not make any sense. When testing the plugin's impact on page load time, the test has to be done in an isolated environment to get results that have real value.

    Moreover, if the response time is so important for your web sites, you can easily take the time, which is required for the generation of the dynamic page, out of the equation altogether by using a caching plugin or a CDN. Especially, if we are talking about public content that is accessed by non authenticated users over the internet, then this discussion is total nonsense. Just use a caching plugin!

    Having arguments on the internet is not my thing either. But it is me who does the development and provides the support, all free of charge. Please allow me to have that 'tone' when I have to spend even more time dealing with posts like yours. If you are having problems with the plugin, please either stop using it or provide me with the necessary feedback, so as to try to fix it. In any case, please stop posting information about a 50-100ms delay in page load time without any proof or instructions about how to reproduce it. This is something I particularly dislike.

    As far as I am concerned this discussion is over. Please post the info I asked for in the WordPress Support forums and I'll see what I can do when I have the time.
     
    George Notaras, Mar 2, 2015 IP
  18. Revelations-Decoder

    Revelations-Decoder Well-Known Member

    Messages:
    3,028
    Likes Received:
    152
    Best Answers:
    4
    Trophy Points:
    190
    #18
    @George Notaras

    I use "Add Meta Tags" and have done for years after trying all the others and I am actually surprised more people don't use it tbh as IMHO it is one the best and most useful WP plugins out there...

    I get page 1 results on Google, Bing etc using it so I am happy with "Add Meta Tags"! :)

    Jeffr2014 sorry for slight tangent but I don't get any loss of SERPS due to Add Meta Tags and WordPress sites fair pretty well compared to other sites but updates can be set to be automated - though I am guessing that would be too risky if you don't trust the plugins you are using.
     
    Last edited: Mar 2, 2015
    Revelations-Decoder, Mar 2, 2015 IP
  19. George Notaras

    George Notaras Greenhorn

    Messages:
    6
    Likes Received:
    1
    Best Answers:
    0
    Trophy Points:
    13
    #19
    Thanks for the kind words! I started developing this plugin over 9 years ago and, despite the fact that I have extremely little free time these days, I intend to continue developing it as long as I have even a single website to maintain. Recently, WordPress.org started displaying the number of active installs for each plugin (instead of the previously displayed number of total downloads). It was a surprise to find out that there are over 100000 active installations, although the plugin has never been promoted or marketed in any way. I guess it does what it says on the tin and people use it. Certainly, it could be better, there is always room for improvement, but development is slow due to the lack of free time.

    About automatic plugin updates and as far as Add-Meta-Tags is concerned, special care is taken that new releases do not break anything. The plugin uses the documented WordPress API in almost all cases. Automatic updates should be safe, but there is no guarantee about it, so not recommended.
     
    George Notaras, Mar 2, 2015 IP
  20. Jeffr2014

    Jeffr2014 Active Member

    Messages:
    254
    Likes Received:
    18
    Best Answers:
    0
    Trophy Points:
    55
    #20
    George, I went through our records and the last time we updated AMT (prior to last update) was Jul. 11, 2014. So the version number in question is the latest version number at that time. It should be easy for you to find the number based on change logs... Our tests with P3 after installing version 2.6.7 (these were conducted by Bill who is the OP at WordPress support forums) reported about 50% increase in time taken by this plugin (with Opengraph, Dublin Core, and schema.org metadata enabled) - according to P3 time increased from 0.016sec to 0.023 sec.

    I am not going to argue with you re validity of P3 vs. webpagetest.org, when I was talking about 0.05-0.1 sec for AMT plugin I was referring to the range of difference reported in 20 webpagetest.org tests (10 with plugin enabled vs 10 with plugin disabled).

    There is not much I can add here, these are all the numbers that we have. We are not in business of evaluating plugins and I regret that by using it as an example in my first post (to illustrate my point about dependency of WP blog owners on plugin and theme developers) I negatively affected AMT reputation - as I repeatedly mentioned here it is one of better plugins out there...
     
    Jeffr2014, Mar 2, 2015 IP