Joomla - Feedback and Ideas For Improving The Work in the Administration In Future Versions For a More Efficient Experience

27 June 2025

Recently, Joomla! 8 - The future is planned now article was published in the Joomla Community Magazine.
It talked about Joomla 8 and the planning process behind its future development.
The article also invited community members to share their feedback and ideas for the upcoming version.

As a developer, I work with Joomla every day.
I hear a lot of feedback from users - what helps them, what slows them down, and what they wish was better.

In this video, I’ll share a few ideas based on real user feedback - simple ways to make the Joomla admin faster and easier to use.

1) Ajax Save → Save Instantly with Ajax

Let’s start with saving items - and we’ll use file editing as an example. Imagine you're editing a file and working on a specific line. You click Save and then switch to the frontend to see how the change looks. But after saving, the editor scrolls back to the top. You lose your place - and you have to scroll back down every single time. It’s a small thing, but it quickly becomes annoying and breaks your focus.

With Ajax saving, that doesn’t happen. The page stays where it is, the save is faster, and you can stay focused on what you’re actually doing. And it’s not just about editing files - the same applies when you're adjusting article settings, editing module parameters, or working with any other content.

2) Ajax Actions → Quick Actions via Ajax

The same thing happens with almost any other action. Every time the page reloads, we lose our focus - and it takes a moment to figure out where we were. With Ajax-based actions, that’s no longer an issue. Everything feels faster and simpler - but more importantly, we stay in control. There’s no interruption, no need to refocus. We can just keep working.

3) Edit in Place → Edit Content Inline

Just like Ajax saving, inline editing can make a big difference. Let’s say we want to quickly change the position of a module. Right now, we have to open the module list, click into the specific module, and change the position manually. If you need to update several modules, it becomes time-consuming and repetitive.

With edit in place, you could simply change the position directly from the module list - no need to open each item. And of course, this applies to many other items that often need quick edits. It’s a small change, but it saves a lot of time and makes the whole process much smoother.

4) Faster Access to Some Sections → Speedier Navigation to Key Sections

Another area that could be improved is faster and clearer access to certain sections. Let’s take the extension installation as an example. Right now, it takes two clicks instead of one - which isn’t a huge issue by itself. But the real challenge is for new users. When they click on the System menu, they see a long list of links to different sections all on one page. This can be overwhelming and confusing for beginners.

It takes them a while to find their way around, simply because there are too many options displayed at once. So this is definitely something to think about - maybe there’s a simpler, more user-friendly way.

5) Unobtrusive, but Still Accessible Help → Subtle, Accessible Help When You Need It

Inline help is another place where small changes could make a big difference. Imagine you find a setting you need to adjust, but there’s no simple explanation nearby. You click the inline help button, but instead of getting a quick description, you lose your focus completely. That breaks your concentration and interrupts your workflow.

This can be solved easily - for example, by showing help text in tooltips. If you’re an experienced administrator, the help won’t get in your way or clutter the form fields. But if you really need guidance, just hover over the setting and the explanation will appear right there. The key is that you don’t lose focus - nothing interrupts your concentration.

Conclusion

These are some really interesting insights that show us how to work more comfortably and efficiently in the Joomla administration - without anything getting in the way or interrupting our flow.

Updated by users

6) Displaying tags in article list

Displaying tags or other relevant information associated with an article in the article list is also an important feature. When browsing through the list of articles, it allows users to quickly get an overview of which articles are linked to specific topics, keywords, or other relevant data. This feature enhances navigation and helps users identify connections between different pieces of content at a glance.

Inkscape - Place Text Behind Object

12 November 2024

This article is a continuation of the previous article: Inkscape - Remove Image Background (AI) Extension.

If we already have a background removal environment for images (rembg library and Inkscape plugin Phoca - Remove Image Background (AI)), then we can simply create other different effects in Inkscape.

One such effect might be to insert text behind an object in a raster image. All can be seen in this video:

The principle is very simple. We make a copy of the main image. On the copy, remove the background according to this tutorial: Phoca - Remove Image Background (AI) and add text. So we will have three layers:

- the main image
- a copy of the image without the background
- text

Now we just need to move the text layer between the main image layer and the copy image layer.

Inkscape - Remove Image Background (AI) Extension

04 November 2024

Phoca - Remove Image Background (AI) extension is an Inkscape extension to help you easily remove backgrounds from raster images (JPG, PNG, WEBP, ...) using AI and the Rembg library.

With Phoca - Inkscape Remove Image Background extension you can remove image background with help of different AI models. There are following options:

• Model - select model
• Post Process Mask - Yes | No
• Alpha Matting - Yes | No
• Alpha Matting Foreground Threshold - set value for Alpha matting foreground threshold (0 - 255)
• Alpha Matting Background Threshold - set value for Alpha matting foreground threshold (0 - 255)
• Alpha Matting Erode Size - set value for Alpha matting erode size (0 - 40)
• Debug

Usage

• Make sure you have your Rembg library environment set up correctly on your computer, see Be aware section below.
• Select raster image, click on Raster - Remove Image Background. Set options and click on Apply button.
• Then the selected image should have its background removed using AI.

Be aware

• This extensions requires Rembg and PIL libraries (and their dependencies) installed on your system, see: https://github.com/danielgatis/rembg
• This extension is for experienced users who have set up and tested the Rembg - a Python library on their computer.
• Removing image backgrounds using AI models can be a very challenging process. Therefore, it is a good idea to have your environment set up correctly and try first to remove the background of an image using the Rembg library directly in the console

How to install this extension

Just follow Inkscape guides for installing Inkscape extensions (download the ZIP package and unzip it to Inkscape extension folder and restart Inkscape)

This extension needs the Python libraries Rembg and PIL and many of their other dependencies to run properly. So before using this extension, you need to install this complete environment using Python.

Optimizing Image Background Removal with AI Models

01 November 2024

In recent years, advancements in artificial intelligence have brought about impressive tools that simplify tasks which once demanded extensive manual labor and expertise. One such application is the use of AI models for background removal in images - a task that traditionally required attention to detail, often done manually with graphic design software. While AI offers significant efficiencies, each model comes with its own strengths and limitations, presenting new challenges and considerations. This article explores these challenges, showcases examples, and demonstrates how combining multiple AI models can yield results that are both more accurate and flexible.

A video is also available for this article, showcasing the processes described.

The Challenge of Imperfect Background Removal with AI

Background removal may seem straightforward, but it often involves complex decisions, such as differentiating between objects and their backgrounds based on color, shape, and texture. AI models use training data to make these decisions, but no single model is perfect for all scenarios. A model that performs well on one type of image may fail on another due to variations in background complexity, lighting, or object color. Therefore, when using a single model, imperfections are almost inevitable. This could manifest as either a part of the object being erroneously removed or the background not being entirely erased, leaving unwanted remnants.

Many professionals who work with these tools know how frustrating it can be to spend hours fine-tuning an AI model’s settings or manually correcting small errors, only to find that the results are still not up to par. Fortunately, combining multiple AI models can offer a solution.

Case Study 1: Combining Models for Optimal Results

To illustrate the benefits of combining models, let’s explore a specific example where each model, when used alone, falls short. In this example, we have two AI models, each with unique strengths and weaknesses:

• Model 1 is proficient at removing backgrounds but occasionally removes parts of the object if it shares a similar color with the background. This can be problematic when the object of interest is not well-delineated from the background, as the model may interpret sections of the object as part of the background and erase them.

• Model 2, on the other hand, tends to leave slight edges or shades around the subject, even though it effectively removes most of the background. While it preserves the main object better than Model 1, these residual edges often require manual post-processing.

By combining these models in sequence, we can overcome their individual limitations. In this example, we use Model 2 as the initial layer of processing, allowing it to remove most of the background while leaving a small edge around the subject. This step ensures that the object remains mostly intact, even if it shares colors with the background. Next, we apply Model 1 to refine the edges, removing the remaining unwanted background and creating a cleaner, more accurate result.

This two-step approach effectively compensates for the shortcomings of each model, delivering a refined image that neither model could produce on its own. This example illustrates how a hybrid approach can reduce manual intervention and produce a polished result more efficiently.

Image Choice: The Key to Successful Background Removal

The effectiveness of a background removal model doesn’t just depend on the AI - it also depends on the characteristics of the input image. Images with high contrast between the subject and the background generally yield better results with single-model processing, as the AI can more easily differentiate between the two. However, real-world images are rarely this straightforward. An object may have colors or textures that are too similar to the background, causing even the best models to struggle with segmentation.

Interestingly, a result that appears unsatisfactory for one purpose may be exactly what is needed for another. For example, leaving a slight edge around a subject may be undesirable for certain professional contexts, but it could work well for marketing materials where the subject stands against a similar-colored background. Likewise, a model that removes parts of the object can be useful in cases where a silhouette or a partially masked effect is desired.

This variability emphasizes the importance of carefully selecting images based on the intended outcome and choosing models or combinations accordingly. In cases where achieving precision is a priority, using multiple models becomes essential to refining the output to the highest possible quality.

Case Study 2: Tailoring AI Models to Specific Requirements

Another practical example highlights the need for flexibility in background removal. Suppose we have an image with an object that shares colors with the background, but in this case, we actually want to remove this object entirely. The first AI model successfully retains the object, assuming it to be part of the primary subject. While this may be useful in some situations, it contradicts our specific requirement in this scenario.

To achieve the desired result, we turn to another model that is better suited to identifying and removing the object completely. This process allows us to eliminate the unnecessary object while preserving the quality and accuracy of the image. This example underscores the adaptability of AI - using multiple models, we can achieve vastly different outcomes based on our needs without being restricted to a single interpretation of the scene.

The Synergy of Combining AI Models

The combined use of multiple AI models in image processing unlocks new potential for more nuanced and reliable results. While individual models may fall short in certain respects, their combination can address complex segmentation issues that would otherwise require intensive manual correction.

Here are some advantages of this approach:

• Increased Accuracy: By layering models with different strengths, it’s possible to address edge cases that would be challenging for any single model, resulting in a more precise background removal.

• Reduced Manual Intervention: With the appropriate model combinations, the need for tedious manual adjustments decreases significantly, saving time and effort.

• Customization for Specific Requirements: Different projects have different requirements. Some might need perfect isolation of the subject, while others might be looking for a more stylized result. Combining models offers the flexibility to tailor outputs based on unique project demands.

• Enhanced Quality: The hybrid approach allows for high-quality outputs that can be used in professional environments where accuracy and clarity are important.

• Improved Efficiency: For high-volume work, such as in e-commerce or media production, using model combinations can create consistent, high-quality results more efficiently than relying on a single model or manual methods.

Conclusion: Embracing the Power of AI Model Combinations

The field of image processing is a dynamic landscape, and AI has undoubtedly become a powerful tool within it. However, it’s essential to recognize that no single AI model can handle all types of background removal with perfect accuracy. By combining multiple AI models, we gain the ability to tailor our approach to each image’s unique characteristics and our specific goals. This synergy minimizes the weaknesses of individual models while amplifying their strengths, ultimately leading to cleaner, more reliable, and more versatile results.

As AI technology continues to evolve, we can expect even greater specialization among models, which will open new doors for creative and practical applications alike. In the meantime, the approach of combining models stands as a valuable strategy, empowering users to harness the full potential of AI in creating polished, professional images. Whether for design, marketing, or other industries, mastering this technique will be a crucial asset for anyone seeking to make the most out of AI-driven image processing.