This commentary does not intend to follow in this footsteps. On the contrary, instead of having endless discussions in the developer world about the rise and fall of individual languages, we should take a look at reality. Here we can see that between all the trends, hypes and supposed extinction, many more languages survive and fulfill a purpose than it appears on the surface. It also shows that programming languages cannot simply be declared superfluous or even dead – at least as we know them today. On the contrary, they are becoming more compatible with each other and are finding specialized use cases.

The currently undisputed Python has experienced its success story with the emergence of AI, data science and machine learning – and has thus moved directly into the fast lane. In view of this meteoric rise, it is easy to forget that the language has been around since 1991 and has not played a serious role in the IT world for decades. Today, it is at the top of the podium and there is no end in sight to its rise. Does this mean we have to retire older languages? Not for a long time yet, as the example of C++ shows time and again: for years, the object-oriented language has been predicted to be doomed, but its numerous use cases, efficiency and versatility ensure that C++, which was developed in the early 1980s, remains stubbornly at the top of the indices.

Programming languages cannot be argued away, the market and the respective communities ultimately have a greater influence than the comments section on the Internet. Last but not least, the widespread legacy applications, in whose depths completely different antiques are still doing their job, are also decisive. The Delphi conference, which is still taking place, will be able to confirm this. Nevertheless, the emergence of new technologies, especially AI, is having a major impact on which languages receive how much attention and which new generations of programmers study at universities. What does this mean for “oldies” like Java, C# or C?

Of course, no one has a digital crystal ball, but a look at the historical development shows that a large number of languages have co-existed for years. Even if Python will continue to be the measure of all things in AI-enthusiastic times, Java, C++ and the like will remain with us for a long time to come – because the use case is there and entire generations of developers will not give up their passion overnight. Looking even further into the future, however, the question arises as to whether our current idea of the programming language concept will endure in the long term. With the further development of AI co-pilots and generative models as interpreters for natural language to assembler, a real turning point would not be unlikely. Why shouldn’t AI one day compile directly in assembler and no longer take the detour via a language such as Python or Java? Perhaps this is why, in the long term, the question is whether programming languages as we know them today will generally be replaced in the future. Until then, many lines of code will still be compiled – in Python, in Java, in C++ and yes, also in Delphi.

This press release is also available at www.pr-com.de/de/avision.

Press contact

Avision GmbH
Christina Karl
Marketing
Bajuwarenring 14
D-82041 Oberhaching
Phone +49-89-623037-967
christina.karl@avision-it.de

www.avision-it.de

PR-COM GmbH
Melissa Gemmrich
Sendlinger-Tor-Platz 6
D-80336 Munich
Phone +49-89-59997-759
melissa.gemmrich@pr-com.de

www.pr-com.de

Der Beitrag Is Python sending Java and co. into retirement for good? erschien zuerst auf Avision.

Does GitHub’s Copilot finally provide developers with an intelligent pair programming partner that both reduces the complexity of the coding world and is an effective means of combating the shortage of skilled workers? To answer this question, there is no way around a practical text. Avision took a close look at the AI tool and put its potential to the test with a focus on TypeScript, Java, React, Python and Angular.

Up-to-dateness and training data

At the time of the test phase, the GitHub Copilot was accessing data from the end of 2021 – although the database is regularly updated, the model is still not up to date. As a result, the latest versions of languages and frameworks such as Angular or React are simply not available. This can quickly lead to further problems, as the tool may not be familiar with new language constructs or security standards, which poses potential risks and can reduce the quality of the code.

Quality of the results

The AI assistant can take over many simple but time-consuming work steps and inspires new approaches to solutions. However, the current version is not a simple copy-and-paste solution. In almost all cases, the human developers had to revise and adapt the generated results. In some cases, Copilot also simply delivers incorrect results that were not recognizable at first glance – this increases complexity and reduces the initial time saved through automation. Nevertheless, the tool is suitable for creating an initial code base on which developers can build. Even if there is still some catching up to do in terms of results, none of the testers would want to do without Copilot.

Potential and productivity

Despite the qualitative gaps and the lack of up-to-dateness, Copilot has enormous potential. We are only at the beginning of a continuous development that will undoubtedly lead to an increase in efficiency and productivity in the future. Until then, however, there is still a lot of work to be done to improve the reliability and accuracy of the results. However, this should not diminish the importance for the industry; AI assistants will be part of the permanent repertoire of developers in the foreseeable future.

Inexperienced developers and generative AI

The question of how developers without many years of coding expertise should deal with generative AI tools is particularly relevant. The risk factor here is that the uncontrolled use of these tools leads to a deterioration in code quality, especially if inexperienced developers tend to adopt the proposed solutions without reflection. On the one hand, this can lead to inconsistent and possibly faulty code that impairs the overall quality of the project. On the other hand, junior developers also lack the necessary learning effect if they only work with AI-generated code suggestions instead of developing it themselves. What leads to a stagnation of the individual learning curve in individual cases can pose a challenge for the promotion and personnel development of the entire industry in the big picture. It is therefore important that inexperienced developers continue to learn through practical experience and do not rely exclusively on generative AI tools.

Conclusion

The test results with GitHub Copilot are mixed. While there are undoubtedly helpful use cases and the tool can automate smaller tasks, developers need to critically scrutinize and review the suggestions generated by the AI at all times. The challenge is that errors or incompleteness are not always recognizable at first glance, which makes using the tool quite complex and time-consuming. Nevertheless, GitHub Copilot offers the promising potential to increase the productivity of developers – at least if the quality and reliability of the results increase. The bottom line is that it is and remains essential to handle the use of these tools responsibly and ensure that they complement and do not replace human expertise.

This press release is also available at www.pr-com.de/de/avision.

Press contact

Avision GmbH
Christina Karl
Marketing
Bajuwarenring 14
D-82041 Oberhaching
Phone +49-89-623037-967
christina.karl@avision-it.de

www.avision-it.de

PR-COM GmbH
Melissa Gemmrich
Sendlinger-Tor-Platz 6
D-80336 Munich
Phone +49-89-59997-759
melissa.gemmrich@pr-com.de

www.pr-com.de

Der Beitrag What can the GitHub Copilot do? erschien zuerst auf Avision.

This press release is also available at www.pr-com.de/de/avision.

Press contact

Avision GmbH
Christina Karl
Marketing
Bajuwarenring 14
D-82041 Oberhaching
Tel. +49-89-623037-967
christina.karl@avision-it.de 

www.avision-it.de     

PR-COM GmbH
Melissa Gemmrich
Sendlinger-Tor-Platz 6
D-80336 München
Tel. +49-89-59997-759
melissa.gemmrich@pr-com.de

www.pr-com.de

Der Beitrag Is the cloud dream already over? erschien zuerst auf Avision.

This press release can also be found at www.pr-com.de/de/avision

Press contact

Avision GmbH
Christina Karl
Marketing
Bajuwarenring 14
D-82041 Oberhaching
Tel. +49-89-623037-967
christina.karl@avision-it.de 

www.avision-it.de     

PR-COM GmbH
Melissa Gemmrich
Sendlinger-Tor-Platz 6
D-80336 München
Tel. +49-89-59997-759
melissa.gemmrich@pr-com.de

www.pr-com.de

Der Beitrag Artificial intelligence, yes but… erschien zuerst auf Avision.

Whether in software development, production or service, solutions from the field of robotic process automation (RPA) and other technologies have experienced a real boom and have become an integral part of most companies’ day-to-day business. Their use ranges from self-running tests for programmers and autonomous production lines in industry to almost human chatbots for customer communication – to name just a few examples. In the world of software development, however, those who do not fully automate all aspects of their work are well advised. Avision uses three key examples to show when human programmers achieve the better results.

The individual design

When it comes to an intuitive user interface that takes the customer’s requirements and needs into account, automated tools are currently reaching their limits. And that’s a good thing, because the human creativity and understanding of employees can make the difference in a competitive market compared to providers who rely on automatically generated solutions. Companies are also more flexible when it comes to unpredictable events such as last-minute change requests if they create the design of user interfaces manually, for example.

The complex use case

Rapidly growing and complicated software can make various aspects of automation more difficult, such as code generation or automated testing. If the complexity of the software increases, the effort required to fully automate it also increases. It makes more sense to place complex functions or algorithms in the hands of human programmers and thus retain more control over the project.

The unnecessary test

Automated code review has many advantages, in particular the rapid detection of errors, the improvement of quality and compliance with standards. However, automated testing is not synonymous with perfect results and cost-effectiveness must also be considered: Anyone who invests 20 days in automation in order to save a one-off five days of manual testing at the end has missed the real target. In these cases, it is still true that people can sometimes be more effective than technology. 

“Automation may break through current boundaries in the near future – what is impossible today may become reality tomorrow,” summarizes Nadine Riederer, CEO at Avision. “Nevertheless, programmers don’t have to automate today or tomorrow just because they can. There will always be cases in which manual, human work offers convincing advantages. The right mix makes all the difference.”

Der Beitrag Everything has an end, even automation erschien zuerst auf Avision.

Germany is to be climate-neutral by 2045, according to the German government’s coalition agreement. A major task that can only be accomplished together. The IT sector also has a role to play here – after all, advancing digitalization is leading to enormous emissions and rising electricity consumption. One component of the strategy: climate-conscious programming for more effective and climate-friendly applications. The IT service provider Avision shows which points can make the difference and what green coding looks like in practice.

Choose efficient data structures

In order to reduce the energy consumption of systems, some data structures are more suitable than others, as they require less computing power and storage space. Examples of this are vectors, linked lists or hash tables. The use of efficient algorithms that are tailored to the requirements can also make a significant difference.

Making the front-end lean

Revising your own website in terms of efficiency has enormous potential for saving resources. Optimizing loading times, for example, is a good way to reduce computing power. Other factors include the media used, such as large images or videos and background processes, which often increase energy requirements unnecessarily. Taken together, these points can significantly reduce the required CPU load and power consumption. In addition, resource-saving and less powerful CPUs can be used.

Use caching techniques

Instead of always updating all values, the use of caching reduces energy consumption. Frequently accessed data is temporarily stored in the cache – when a new request is made, users have faster access to it and do not have to send a new request to the server. The decreasing number of updates reduces the required workload and network traffic. There are different techniques for different requirements, such as client caching, server caching or CDN caching.

Using cloud computing and virtual machines (VM)

The major advantage of cloud computing and virtual machines is their scalability. They can be adapted to the required computing power and users can, for example, shut down test systems directly in order to save money and electricity. With dedicated hardware, on the other hand, frequent shutdowns can lead to damage. Running multiple applications on a single virtual machine instead of multiple physical devices also helps to save power. Last but not least, VMs also help to reduce the amount of space required on the hardware. At this point, it is important that software development already takes these scenarios, especially scaling, into account when coding.

“Well-maintained and efficient code can contribute much more to saving energy than it seems at first glance – a factor that will hopefully be reflected in more and more project plans in the future,” says Nadine Riederer, CEO at Avision. “A shining example of this is the cryptocurrency Ethereum, which was able to reduce its own energy consumption by 90 percent with code adaptations and a proof-of-stake algorithm. Green coding has enormous potential.”

This press release can also be found at www.pr-com.de/de/avision

Press contact

Avision GmbH
Christina Karl
Marketing
Bajuwarenring 14
D-82041 Oberhaching
Tel. +49-89-623037-967
christina.karl@avision-it.de 

www.avision-it.de     

PR-COM GmbH
Melissa Gemmrich
Sendlinger-Tor-Platz 6
D-80336 München
Tel. +49-89-59997-759
melissa.gemmrich@pr-com.de

www.pr-com.de

Der Beitrag Programmed sustainability: implementing green coding correctly erschien zuerst auf Avision.

First of all: of course, every language has its right to exist and useful use cases. It is also important, especially in the fast-moving and innovative IT sector, to keep new trends on the radar, to educate yourself further and not to dismiss developments on the market as newfangled stuff per se. However, and this should also be factored into any decision-making process, not every trend reinvents the wheel. It’s more likely to be the old hands such as Java, C, JavaScript or PHP that continue to provide reliable service, receive the necessary support and have enough knowledgeable developers. Because even the most innovative project that has been realized with the most unusual new language is dependent on people who will still be able to read the code in the future. If this is no longer the case, companies quickly face a huge problem -NASA can tell you a thing or two about it. As important as progress is, for business-critical projects that are designed to run for many years, the choice of Java and co. provides better odds when betting on which languages will still be popular in the future. The rise of Python impressively demonstrates that these developments are by no means set in stone. While the language had a shadowy existence for a long time, the emergence of AI, among other things, helped Python to become one of the most popular and widespread programming languages of all.

As we all know, life is not a wish list and companies are often faced with the challenge of implementing projects in the shortest possible time. Those who buy into the promises of many newer tools, frameworks and languages for even faster results may actually be able to bring products to market and applications into production faster – but at what cost? What if the highly praised new framework is so irrelevant after three years that no developer bothers to learn it anymore? What if the new language not only loses its status in the TIOBE index, but also no new versions are released that should actually close the security gaps that are increasingly being discovered? 

The choice of programming language is and remains a bet. Established languages clearly have the better odds when it comes to long-term and secure planning – this is by no means to demonize innovation. We just don’t have to chase every trend in order to achieve the best possible results. Software development is no exception.

Der Beitrag Programmers, stick to your languages erschien zuerst auf Avision.

Of course, software developers also make mistakes. Some of them are even particularly widespread. To prevent them from spreading further, IT service provider Avision explains some extremely popular blunders in development projects:

Because they are simply nice and the product owner or the specialist department want it so much, developers add a new requirement shortly before the end of the sprint. This is not a good idea, because in the end it is precisely their implementation that causes an error. This happens particularly often with very small requirements that seem unproblematic and are therefore not really taken seriously.

Another popular last-minute mistake: replacing a component shortly before the end of the development project. A bug has cropped up in a framework that we use? Then we just use a newer version of it. Unfortunately, it turns out that this one has five other bugs.

“I made a quick adjustment because it wasn’t nice.” Sounds like commendable initiative, but all too often it backfires. With such actions, developers usually do not write test cases and do not check what effects their adjustments have on the rest of the software.

Is the development project behind schedule? Then we simply switch from a conservative to an agile approach – after all, agility is a guarantee for rapid development. Unfortunately, this is not true. Agile development means meeting the needs of users better because you are more flexible. Agility ensures that you hit the target more precisely, not faster.

Another common attempt to make up for lost time is to increase the size of the team. This can even work if you increase the team from ten to eleven employees. But five new employees are too much of a good thing. They all have to be trained and the communication effort within the team increases significantly. What’s more, some processes simply take a certain amount of time – no matter how many people are working on them.

Supposed time saving for the third: In order to be faster, the team increasingly relies on test automation when testing the software. However, the situation here is similar to that of agility. Automated tests make the code better, but it is not finished any faster. As the teams have to develop the test cases in addition to the code, they do not make any faster progress overall.

“You learn from your mistakes. And you become particularly wise when you learn not only from your own mistakes, but also from the mistakes of others,” says Nadine Riederer, CEO of Avision. “Sometimes, even in development projects, you can’t avoid paying for lessons learned, but it’s only sensible to keep these expenses as low as possible.”

Der Beitrag Development teams should not repeat these mistakes erschien zuerst auf Avision.

However, this does not mean that developers can simply ignore software testing, says IT service provider Avision. On the contrary: quality cannot be “tested into” software retrospectively. That’s why developers need to consider the issue from the outset and provide the best possible support for the testers’ work.

In particular, you should:

• always consider testing when developing their software and program options for this; ideally in the form of automated tests.

• understand the technical requirements for their software. Only then can they write meaningful tests for their components. To develop this understanding, they need to communicate: with the project management, the specialist departments and the testers.

• always bear in mind that users might use the software differently than intended. You should be open to user errors and also take possible incorrect entries into account during development.

• do not see testers as opponents, but as partners who support high software quality. The opponents are the errors in the code – and not the testers who find them.

“Developers should have high quality standards for their software and discover standard errors themselves,” says Nadine Riederer, CEO of Avision. “It is then the testers’ job to think outside the box and carry out unusual tests. They can also put the big picture to the test, which developers often can’t see because they only create individual components.”

Der Beitrag Why software testing depends on the developers erschien zuerst auf Avision.