<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Dimitris Grammatiko… on Medium]]></title>
        <description><![CDATA[Stories by Dimitris Grammatiko… on Medium]]></description>
        <link>https://medium.com/@dgrammatiko?source=rss-5767bcc06d82------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*m78LLyPSu5oUQtMJ.jpeg</url>
            <title>Stories by Dimitris Grammatiko… on Medium</title>
            <link>https://medium.com/@dgrammatiko?source=rss-5767bcc06d82------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 05:04:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dgrammatiko/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Sorry but we are doing it {awfully} wrong]]></title>
            <link>https://medium.com/@dgrammatiko/sorry-but-we-are-doing-it-awfully-wrong-910c0f0dbbb0?source=rss-5767bcc06d82------2</link>
            <guid isPermaLink="false">https://medium.com/p/910c0f0dbbb0</guid>
            <category><![CDATA[web-development]]></category>
            <dc:creator><![CDATA[Dimitris Grammatiko…]]></dc:creator>
            <pubDate>Fri, 17 Aug 2018 16:11:14 GMT</pubDate>
            <atom:updated>2018-08-18T14:12:12.713Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_LP5NU_Jq-JJwkxTm0kL7Q.png" /></figure><p>So there are some Joomla contributors that are preparing the next major version of the popular CMS. Unfortunately, there is a decision that is totally wrong and there will be a really heavy price that the project will pay in the future because of this…</p><h3>Context</h3><p>Decisions without any valid context mean nothing, so let’s define the context first of all. Joomla 4 will be the successor of version 3 and fundamentally the project shouldn’t try to lose any of its current users due to a difficult upgrade path. This, according to the available information, will be a “one button click”.</p><p><strong>Right, and I’m an astronaut!</strong></p><p>So the real story is that this will be true <strong>ONLY</strong> if you keep your old Bootstrap 2.3.2 templates (assuming that you’re using Joomla the right way and you already created layout overrides for everything). If you’ve invested time to create any overrides, all the pros rightfully did so, then you face a huge problem: you need to rewrite everything to Bootstrap 4 (assuming that components, modules and plugins exist for the new version). Now that’s a little bit more than one button click for me…</p><p><strong>But wait it get’s better…</strong></p><p>So if for whatever reasons, you’ve decided to actually go ahead and redo all the overrides to satisfy the new requirement, that is Bootstrap 4, there is no guarantee or whatsoever that this will not be happening in a very short period. Why do I say this? Well, there is already a version 5 tag for Bootstrap (and although it supposed to be only the removal of jQuery there is little assurance that nothing else will be broken). But even if Bootstrap version 5 is 100% B/C (yeah, well 😂) there is a far greater picture that people tend to miss. Bootstrap is already kinda deprecated because the platform itself provides much more powerful and future-proof solutions, like CSS Grid Layouts and of course CSS Variables (or custom properties). Funny stat here is that these technologies have the exact same minimum requirements as Bootstrap 4!!!</p><p><strong>But wait it gets even better…</strong></p><p>After two decades that programmers tested various technics and solutions on the web, there is some hard gained knowledge which might be presented as buzzwords like PWAs, etc. In reality what happened the last 5 years or so, is that all the major browsers sat repeatedly on the same table discussed various approaches and finally came up with the ultimate solutions. CSS Grid Layouts, Web Components, Javascript ES6, CSS Houdini, WASM, etc… All these technologies have been pushed almost simultaneously across the browsers (check caniuse.com, you’ll be surprised). So when people had a verdict that the first response needs to had a critical inline CSS and then lazy load the fancier part of the page’s stylesheet, it’s not f@#$ing buzzword that you might say Nah… But actually you need to say NAH because you decided to force Bootstrap and Bootstrap won’t allow you to sensibly split the CSS unless you totally rewrite it from scratch!!!</p><p><strong>But wait it gets even better…</strong></p><p>If you know one or two things about Joomla’s architecture then you already know that everything originates from a component’s view. The new router by Hannes was built upon this very fundamental concept. So every component view is responsible for a page rendering, modules are attached to that view and of course plugins are executed in the application’s lifecycle. So Joomla is already modularised to the greatest possible extent! Now how the f@#$k did we end up from a totally modularised system (that fits perfectly the today’s needs) to have a GLOBAL CSS file that controls all the views is something that historians could try to explain. So Joomla is, in fact, a perfect base for a modern website (have you seen my presentation <a href="http://twitter.com/JAB">@JAB</a> 2018, if not you should) the problem is that people WANT to DICTATE something that is</p><ul><li>against the systems architecture</li><li>degrades enormously the performance</li><li>creates a lock-in ecosystem, which might throwback sooner or later</li></ul><p>But there is that one answer that really makes me furious:<br><strong>Styling is hard, thus we’re using Bootstrap</strong>. <br>3U!!5H!T. The backend got exactly 3 views: item edit, list view and the landing page. Front-end has one view more (the item view), but honestly, it is not even about quantity it’s all about the misconception here: “Bootstrap is easier”. But hey, I’ve got news for you: if you cannot style 3-pages with plain css then I’m afraid that you are using the wrong tool, you shouldn’t use Joomla you should be using WIX or something similar…</p><p>By the way, if someone is comfortable with the bloated, non performant code provided by Bootstrap, they will still be able to use it, but Joomla out of the box DOESN’T need it and shouldn’t force i!. And that is what this post is all about: core goes lean with extraordinary performance out of the box without restricting people that still want to use something like Bootstrap or jQuery. <strong>Joomla should **not** force any tools either on javascript or on css.</strong></p><p>All and all: Joomla has a great opportunity to go with the tide and modularise everything or try again to swim against it. People should learn their lessons and don’t repeat history because it won’t be a joke. Nope, that’s wrong, actually it will be a joke.</p><p>PS: Mind the road ahead…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8mqSqEtHfskqhgUaSCY6EQ.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=910c0f0dbbb0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Let’s build it then]]></title>
            <link>https://medium.com/@dgrammatiko/lets-build-it-then-ca6541b2cc93?source=rss-5767bcc06d82------2</link>
            <guid isPermaLink="false">https://medium.com/p/ca6541b2cc93</guid>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[php]]></category>
            <dc:creator><![CDATA[Dimitris Grammatiko…]]></dc:creator>
            <pubDate>Wed, 29 Mar 2017 19:00:26 GMT</pubDate>
            <atom:updated>2017-03-29T19:00:26.306Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PUC-j0GnVwETO8rRpJfQzw.jpeg" /></figure><p>An RFC for the GUI override creator</p><p>Typical lifecycle of a get request in Joomla, for the shake of simplicity let’s assume that the template has only one position for rendering the component and there are no modules plugins etc…</p><p>The first computation part till the view call is irrelevant to our example here so let’s assume that we call a single article, the controller and the model do their thing and we are at the point where the view comes to play. The typical code of Joomla for this part is: check if override exists in the active template, if so use that file else fallback to the component/views/viewname/tmpl/default.php. Let’s add one more simple step to this logic, if in the template override path a file named default.tmpl.json exists then we will use a helper file that will parse the default.tmpl.json and load the relative blocks that in essence are just JLayout files.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/28533944a5bd1b4b5adbd108dc220209/href">https://medium.com/media/28533944a5bd1b4b5adbd108dc220209/href</a></iframe><p>So what is in a default.tmpl.json you might ask. Let’s go back and see how that file is getting generated in the first place. The <strong>only requirement </strong>of this system is that each component, module, plugin will provide a description file for every view that is exposing to Joomla. That file could be a simple JSON file, that describes the object that will be passed to the view in every detail <strong>and </strong>the related blocks that each node could use. For example the typical object of an article is an $item which will have a title and that title is bound to the blocks title (in essence the blocks are a folder that contains all the possible variations for the element that needs to be rendered).</p><p>Typical structure of the default.description.json:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/9fd18087600e81cec90cc3a0b9262638/href">https://medium.com/media/9fd18087600e81cec90cc3a0b9262638/href</a></iframe><p>So parsing this file the GUI override builder will get the key and bind it to $item so it will have that $item-&gt;title and the value part is what JLayouts should it be able to offer to the user for this data (the title). User pick whatever they want form the lego box and drop then in the page area. Onve the block is set in the page area more options are vailable to the user, like adding classes etc… Repeating this process till they satisfy their needs and press save and then is the point where the file default.tmpl.json is generated in the template/html/com_content/article/default.tmpl.json. How different is that file from the description? not much. instead of the folder/blocks name it actually have the actuall JLayout paths to the needed layouts and of course the extra key values like class =&gt; “my-nice-class-name”.</p><p>That’s it basically! A simple JSON file for each component/module/plugin that describes the model output and the possible blocks that can be used, and then the generated JSON in the template overrides folder (this can be a new folder instead of the html, needs to be decided I guess).</p><p>GUI builder parses the component/module/plugin JSON and generates another JSON.</p><p>The renderer picks up the JSON in the overrides folder and composes the layout in run time out of the JLayouts.</p><p>Ultra simple approach, should add very little overhead to the current system but will also:</p><ol><li>Give amazing flexibility to end users, because let’s face it, users like to click not to write!</li><li>Decouple once and forever the system from any css framework, without breaking anything!</li><li>Give Front End Devs the ability to work with tools that fit <strong>their </strong>way of working instead impossing Bootstrap or whatever</li><li>Give Component/module/plugin developers the assurance that they won’t have to fight with their customers over their code not behaving well in the x-y-z template, css framework.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ca6541b2cc93" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Solving Joomla’s UX problems]]></title>
            <link>https://medium.com/@dgrammatiko/solving-joomlas-ux-problems-88bcc8a0690c?source=rss-5767bcc06d82------2</link>
            <guid isPermaLink="false">https://medium.com/p/88bcc8a0690c</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[joomla-development]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[cms]]></category>
            <category><![CDATA[joomla]]></category>
            <dc:creator><![CDATA[Dimitris Grammatiko…]]></dc:creator>
            <pubDate>Wed, 29 Mar 2017 13:13:44 GMT</pubDate>
            <atom:updated>2017-03-29T15:03:47.812Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FFUvJgXDH3wxPtT7jM2DTw.jpeg" /><figcaption>Photo: <a href="https://unsplash.com/@ryanjohnstonco">https://unsplash.com/@ryanjohnstonco</a></figcaption></figure><p>Joomla 4.0 is painstakingly crafted by a bunch of people. One of the main areas of improvement was to simplify the convoluted User Experience. So how things are evolving? What’s the plan there?</p><p>Joomla is one of three major open source CMSs, which tends to loose users due to the poor interface and the extremely difficult workflows. A very reasonable reason, since no-one (except hardcore mazochists) wants a system that is extremelly weird and needs some deep knowledge to even complete the very simple tasks. But Joomla used to be one of the pioneers in the Content Management a decade ago, so how that system was ruined so badly?</p><h3>The “Always be yourself” mantra</h3><p>Wordpress and Drupal, right from the beginning of their existence, followed a different path than the one the Mambo (the predecessor of Joomla) chose. The architecture of these systems is so vastly different, so any idea that any of these systems could mirror major the other one is damn stupid! To give you a simple example think of Wordpress as an offroad car and Joomla as a lake boat. You can add wheels to the boat but it will never be a <strong>real</strong> car, or you can add some inflatable balloons in the car but it will never be a <strong>real</strong> boat. So it’s better to understand each system for what it trully is and use (or improve) it as such. This is by far the best course of action. Use each system for what it truly is, don’t try to cannibalize it for whatever stupid reason.</p><blockquote>Well Joomla has been cannibalised to the extreme to match the other two major CMSs, with poor results, because in the process people canceled out the principle “form follows function”</blockquote><h3>Form follows function</h3><p>If you studied design (by the way due to my engineering background I have), you are already familliar with this essential principle. For those not familliar with, let me call it <strong><em>the law, </em></strong>please<strong><em> </em></strong>spent some time to educate yourself, starting here: <a href="https://en.wikipedia.org/wiki/Form_follows_function">https://en.wikipedia.org/wiki/Form_follows_function</a>. So, you might ask, how the heck an industrial, architecture principle is affecting a softwares program? The answer is obvious, or it should be at least for the so called developers, buildings and softwares share something common: <strong>Architecture</strong>. But you already knew that, didn’t you?</p><p>What I’m trying to highlight here (maybe spending more time than what is appropriate) is that the three well known CMSs don’t share the same architecture and therefore will never, ever have the same workflows. No matter what gimmicks and shortcuts and other fancy things you throw at them they can only pretend to be something that are not truly are.</p><p>And that is the problem of Joomla: people want the system to become a clone of Wordpress. They don’t like to put it straight forward like that so they came up with a nice wrapper for that: we want Joomla to have the good UX of Wordpress. The problems start to arise when you blindly copying the workflow of something vastly different, thinking that:</p><blockquote>if it works for that system then most probably it will work for us</blockquote><h3>Hell no!</h3><p>Yes programmers, in essence, most of the times craft code that replicates, or improves a functionality that some other software already introduced. The real “<strong>new</strong>” and “<strong>unseen”</strong> features are vary rare and most of the times those need some time to be accepted by the majority of developers. What I’m trying to highlight here is that you can implement “your” version of drag and drop to your system and that will be a good “copying” of that feature that another system already had implemented. But there is a caveat here, you as developer have the responsibility to adjust that feature to fit the system and not change the system to fit the feature. And there we go, we reached the root of all UX problems in Joomla:</p><blockquote>New features should embrace the system for what it truly is and NOT try to change it</blockquote><h3>Combine and concur</h3><p>So if someone want to distil this post to a couple lines those will be:</p><p>Form always follows function and all features should embrace the system for what it truly is without ever trying to change it’s basic architecture.</p><p>Then someone could ask: is Joomla built uppon a good architecture?</p><p>Oh, yes! Joomla is using the very familiar tree like organization of the data, visualise it as something similar to the filesystem that your operating system is using. Folders can have subfolders (categories and subcategories) and files (articles, banners, contacts, items, etc). Also this structure is already familiar to all the people using a computer as this data structure is apparent to all operating systems (exception IOS, that very cleverly hides it for the end user, but it’s still there). So, a naïve person could think how the heck a so familliar architecture became so painful for users in Joomla? Well the answer is that by time a gazillion features were added and those features were actually fighting the systemby time over time instead of embracing it and improving it.</p><h3>Take a step back and start removing the clutter</h3><p>I didn’t expect that you make till here, but good news as in this section I’ll try to give you my view of an ultra simplified workflow and therefore a vastly improved user experience. Let’s take com_content as our point of reference and particularly the edit view:</p><p>So what will happen if instead of all the known tabs:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/706/1*PYhktUFCj5L5IrOImorp5A.png" /></figure><p>we end up only with the essential ones:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/385/1*lqEmYm7674JqctOwEXEXHQ.png" /></figure><p>Someone could just say “you cannot do that, we’re gonna loose all the flexibility for the displayed article and the ability to control the permissions”</p><p><strong><em>AND THEY’LL BE F@#$NG WRONG, </em>probably, because they don’t understand Joomla! 😮. </strong>Remember all these things I wrote above about the tree like architecture? Well that in Joomla comes with another magic word: “inheritance”. Put it simple: all articles under a specific category should have the same properties which are inherited from the category or the menu, if that article is published under a menu item. But my idea goes a little bit further than that. So instead moving those options up to the tree, which obviously will be a great improvement, why not eliminating them all together and leave the display options to the template, as the architecture already dictates. So my complete idea is: have just a layout selector in the article edit (will default to the one that the category should imply, if none was selected) and have a GUI override creator in the template, with a very clear path that user could, effortlessly create the desired design (without even knowing PHP, HTML, JAVASCRIPT or CSS).</p><p>Does that seems very hard to achieve? Then try asking if people want green or red button to click as they go away…</p><p>EMBRACE WHAT WE FUCKING ARE STOP DRESSING IT UP AS A CLONE</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=88bcc8a0690c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>