With the introduction of native code caching, package authors can now utilize precompile statements or workloads with PrecompileTools in order to cache critical routines earlier on. Users also have the ability to build custom local “Startup” packages that load dependencies and precompile workloads specifically for their day-to-day work. 

However, the company stated that with this capability comes an increase in precompilation time by ten to fifteen percent, and cache files have grown because of the storage of more data and the use of a different serialization format.

Next, the introduction of package extensions automatically loads a module when a set of packages are loaded. The module is contained in a file in the ext directory of the parent package and loads the “weak dependency” and extend methods.

This feature is intended to reduce the amount of capabilities a customer is paying for that they do not actually use. Package extensions also allow for the precompilation of conditional code and the addition of capability constraints on weak dependencies.

Julia 1.9 also brings heap snapshots that can be examined using Chrome DevTools. To create a heap snapshot, the user must use the Profile package and call the take_heap_snapshot function. 

Additionally, to simplify the process of identifying the total number of objects retained, customers can utilize the all_one=true argument and every object’s size will be reported as one, so they can focus more on the number of objects. 

A new command flag had also been introduced, –heap-size-hint=<size>. With this, users can set a limit on memory usage and the garbage collector will then work harder to clean up memory that is not used.

“Julia is incredibly attractive for people who have really demanding computational needs,” said Tim Holy, professor of neuroscience at Washington University School of Medicine and one of the core contributors behind the update. “You can think of previous versions of Julia like a jet airplane: a fantastic way to travel long distances, but probably not your preferred vehicle for going to the grocery store. Julia’s speed on big tasks arises from its ability to generate high-quality machine code; but generating such code takes time, and for a simple task it could mean spending more time generating code than running it. It’s a bit unfortunate for a new user, who has heard so much about Julia’s speed, to have their first experience be waiting for it to compile lots of package code before they can start doing anything useful. No matter how fast it runs once it’s ready, that’s not the best first impression.”

To read the full list of features, visit the website.

The post Julia 1.9 delivers native code caching appeared first on SD Times.

Among these announcements was the launch of the Watch Face Format. This offers developers a new way to build watch faces for Wear OS smartwatches with a declarative XML format to design the appearance and behavior of watch faces.

According to Google, this means there is no need for the executable code involved in building a watch face, and there is also no code embedded in the watch face APK.  With this release, users can create analogue and digital watch faces, watch faces with complications, customizable watch faces, and more. 

Updates have also been made to Jetpack Compose, including the ability to bring Compose everywhere that the developer is building UI, whether that be on a phone, tablet, foldable device, watch, or TV. These changes are available in the May 2023 release and the most recent alpha versions of the Compose libraries.

Furthermore, the performance of Compose has been improved. The company stated that since October, modifiers have started to migrate to an improved system, and with this, text has seen a 22% performance gain that can be seen in the latest alpha release.

The alpha version of the UI framework for creating functional apps for Android TV Compose for TV has also been released. With this, apps can be built with less code, easier maintenance, and a modern Material 3 look. 

Next, Android Studio Hedgehog was launched in order to offer developers an AI powered conversational experience geared at heightened productivity. This offers improved tools when developing for multiple form factors, and helps to improve the quality of applications with new insights, debugging, and testing solutions.

Android for Cars also received updates intended to expand the reach developers have with users in the car.

Among these updates is the global rollout of Waze on the Google Play Store for all cars with Google built-in. Additional templates were also created for Android for Cars App Library in order to accelerate development time across several app categories, such as navigation.

This update also brings tools to port media apps across Android for Cars; new app categories for driving experiences; and video, gaming, and browsing in cars. 

Additionally, Google Play has been updated so that on top of the ability to tailor by country and pre-registration status, users can now customize listings for inactive users, so they have the chance to try out an app or game again.

Google also announced that custom store listing groups for Google Ads App campaign ad groups are set to be launched in the near future. This will allow for custom listings to be served to users coming from specific ads on AdMob and YouTube.

Next, Google Workspace adds Duet AI, combining all of Workspace’s generative AI features and allowing users to collaborate with AI so they can enhance their productivity. Duet AI is also coming to Google Cloud, enabling users to solve common work challenges with the AI collaborator.

There have also been several updates to Vertex AI, including three new foundation models that can be accessed through API, tuned through an AI in Generative AI Studio, or deployed to a data science notebook. 

Lastly, Google Cloud has launched the next-generation A3 GPU supercomputer. With this, Google Cloud offers users a complete range of GPU for the training and inference of machine learning models. 

For more information on these releases, visit the website. 

The post Google announces updates to Android, Google Cloud, Workspaces, Google Play, and more at Google I/O appeared first on SD Times.

IBM is offering an AI development studio with access to IBM-curated and trained foundation models and open-source models, access to a data store so users can gather and cleanse training and tuning data, and a toolkit for governance of AI aimed at making AI more easy to adapt and scale.

The IBM watsonx platform is made up of three product sets: IBM watsonx.ai, IBM watsonx.data, and IBM watsonx.governance. All of these product sets are geared at addressing the need for businesses to build their own AI models, or fine-tune available models, on their own data and deploy them at scale.

“With the development of foundation models, AI for business is more powerful than ever,” said Arvind Krishna, chairman and CEO of IBM. “Foundation models make deploying AI significantly more scalable, affordable, and efficient. We built IBM watsonx for the needs of enterprises, so that clients can be more than just users, they can become AI advantaged. With IBM watsonx, clients can quickly train and deploy custom AI capabilities across their entire business, all while retaining full control of their data.”

IBM watsonx.ai is expected to be available in July 2023 and provides users with a next-generation enterprise studio that allows AI builders to train, test, tune, and deploy traditional machine learning as well as newer generative AI capabilities through an open user interface.

Next, IBM watsonx.data, also expected in July 2023, is a fit-for-purpose data store built on open lakehouse that is intended for governed data and AI workloads, supported by querying, governance, and open data formats to access and share data.

Lastly, IBM watsonx.governance can be expected later this year, and it is an AI governance toolkit that is geared at enabling trusted AI workflows. It operationalizes governance to help reduce the risk, time, and cost that comes with manual processes, and offers capabilities to protect user privacy, detect model bias and drift, and help organizations meet ethics standards. 

IBM also plans to integrate watsonx.ai foundation models into all its major software products going forward, including:

• Watson Code Assistant: A solution that accesses generative AI to allow for the generation of code with an English language command.
• AIOps Insights: AIOps features with foundation models for code and NLP to offer heightened visibility into performance across IT environments.
• Watson Assistant and Watson Orchestrate: IBM’s digital labor products which are expected to be combined with an NLP foundation model to help improve employee productivity as well as user experiences 
• Environmental Intelligence Suite: IBM EIS Builder Edition is intended to allow organizations to create solutions to address and mitigate environmental risks based on their own goals.

To learn more, visit the website. 

The post IBM releases watsonx platform to power next-generation foundation models for AI appeared first on SD Times.

Frappe makes defining DocTypes easier without the need for code to be used everywhere, since everything in Frappe is a DocType. The framework also includes a single-page application with forms, lists, and search navigation built in.

Additionally, users have the ability to define customer roles and permissions that are applied out of the box on every interaction, and they can create their own applications that can be extended by other apps with Frappe’s modular architecture. 

The framework allows for integrations and comes with REST API and Webhooks on all models based on authentication. It also allows users to configure background workers as well as run periodic tasks that are powered by Python RQ.

Soket.io support based on NodeJS is also included, along with the ability to send, receive, view, and manage emails using SMTP and IMAP based email accounts.

Lastly, Frappe enables customers to host several sites on one server, with its database driven multi-tenant architecture. 

The post SD Times Open-Source Project of the Week: Frappe appeared first on SD Times.

Included in this release is a developer preview of a new reactivity model for Angular which provides improvements to both performance and developer experience. This preview is backward compatible and interoperable with the current system.

With the developer preview, users gain access to improved run time performance through a reduction of the number of computations that take place during change detection. It also provides a more simple mental model for reactivity which is geared at making it clear what the dependencies of the view are and what the flow of data through the app is.

Furthermore, the preview enables fine-grained reactivity, makes Zone.js optional in future releases, delivers computed properties without the need for recomputation in each change detection cycle, and offers better interoperability with RxJS by outlining a plan to introduce reactive inputs.

Angular 16 also brings improvements to server-side rendering for Angular with a developer preview of full app non-destructive hydration. With this, Angular no longer re-renders the application from scratch. 

The framework now looks up existing DOM nodes while constructing internal data structures and attaches event listeners to those nodes.

This offers users multiple benefits, including the absence of content flickering on a page for end users, better Web Core Vitals in certain scenarios, future-proof architecture to allow for fine-grained code loading, the ability to integrate with existing apps in a few lines of code, and the incremental adoption of hydration with the ngSkipHydration attribute in templates for components performing manual DOM manipulation.

Angular 16 also offers migration schematics and a standalone migration guide to support developers as they move their apps to standalone APIs.

Developers can access this by running “ng generate @angular/core:standalaone” and the schematics will convert the code and remove extra NgModules classes, changing the bootstrap of the project to utilize standalone APIs.

Users also gain access to several new features with this release, such as: 

• The ability to mark an input as required 
• The ability to pass router data as component inputs
• CSP support for inline-styles 
• A more flexible ngOnDestroy 
• The option to use self-closing tags for components in Angular templates 

To learn more, read the blog post. 

The post Angular 16 offers several improvements to reactivity, server-side rendering, and tooling appeared first on SD Times.

This version also brings multiple other enhancements, including improvements to the way that Upgrade Assistant updates NuGet packages, upgraded Incremental scenario to use YARP 2.0, and improved error handling that shows all failures and warnings in the Progress View for each project component. 

According to Microsoft, in the last version of Upgrade Assistant, when the user chose to upgrade from .NET Core or later to .NET 6, 7, or 8, only the target framework was upgraded. With this release, however, all packages the application is referencing will be upgraded to a set of packages that correspond to the target .NET.

Users can also upgrade existing Xamarin.Forms applications to .NET MAUI, Xamarin’s successor, offering several new benefits and improvements such as single project to simplify asset management; NuGet management; and leverage multi-targeting.

Additionally, this upgrade brings multi-window support for desktop and table scenarios, a rebuilt layout geared at improving maintainability and performance, App Builder to standardize app bootstrapping with common .NET pattern, and a decoupled platform from  cross-platform controls.

There have also been upgrades made to Azure Functions. The company stated that when Azure Functions are upgraded to the latest .NET, the tool will also upgrade the user’s version of Azure Functions to the latest version (v4). The body of functions has also been updated to use the new APIs.

To learn more, read the blog post. 

The post Microsoft’s .NET Upgrade Assistant now offers support for .NET MAUI and Azure functions appeared first on SD Times.

This capability integrates into mabl’s SaaS platform so that users can enhance the value of existing functional tests, move performance testing to an earlier phase of the development lifecycle, and cut down on infrastructure and operations costs.

“The primary goal is to help customers test application changes under production load before they release them so that they can detect any new bottlenecks or things that they would have experienced as the changes hit production before release,” said Dan Belcher, co-founder of mabl.

According to the company, these API load testing capabilities allow for the unification of functional and non-functional testing by utilizing functional API tests for performance and importing Postman Collections to cut down on the time it takes to create tests. 

Mabl also stated that this performance testing lowers the barrier to a sustainable and collaborative performance testing practice, even for teams that do not have dedicated performance testers or specific performance testing tools. 

“Anyone within the software team can use it, so it is not limited to just the software developers or just the performance experts,” Belcher said. “Because we’re low-code and already handling the functional testing, it makes it super easy for the teams to be able to define and execute performance tests on their own without required specialized skills.”

Furthermore, these tests can also be configured to run alongside functional tests on demand, on a schedule, or as a part of CI/CD pipelines. 

The post Mabl’s load testing offering provides increased insight into app performance appeared first on SD Times.

This tool is designed to replace legacy testing grids and add AI-powered self healing capabilities that work towards augmenting open-source test frameworks. These capabilities heal broken tests intelligently as they run, which reduces flakiness and execution time. 

According to the company, this release locates and solves issues continuously, so the user can be sure that tests are running smoothly and reliably. 

It is also intended to be scalable and affordable so that testing teams can test as much or as little as necessary, without the need to worry about bottlenecks or performance issues.

“Applitools’ self-healing Execution Cloud is a game-changing solution that will revolutionize the way companies of any size test their software at scale,” said Adam Carmi, co-founder and CTO of Applitools. “Teams that use open-source test frameworks can now execute their tests in the cloud at scale and take advantage of the latest and greatest AI capabilities that ensure their tests are automatically healed, saving valuable time and effort.”

Applitools Execution Cloud is also compatible with popular open-source frameworks such as Selenium and WebdriverIO. It can also run existing tests without the need for code changes.

To sign up, visit the website. 

The post Applitools Execution Cloud offers users new self-healing capabilities appeared first on SD Times.

At its inception in 2021, FAIR applied recent advancements in computer vision towards an intuitive pipeline in order to animate any human-like figures found in character drawings. 

The demo for Animated Drawings came as a publicly-accessible and web-based tool that allowed users to upload drawings, view/correct model predictions, and receive an animation starring the character in their drawings. 

With the demo, parents also had the option to opt to allow Meta to retain the image and annotations for suture research use. The demo was released with the hopes of collecting 10,000 drawings that could be used to improve model performance.

Today, the open-source version of the Animated Drawings project includes both the models and the code used by the Animated Drawings Demo. 

The company stated that the idea behind the open-sourcing of the project is to enable other developers to create demos and experiences that utilize the animation assets from user drawings. 

This project is also intended to lower the barrier to entry for people who wish to experiment with animation and create their own drawing-to-animation experiences. The release of both the model and code is geared at offering developers a starting point to build on and further extend the project.

To learn more, visit the website. 

The post SD Times Open Source Project of the Week: Animated Drawings appeared first on SD Times.

These features integrate with the company’s no-code visual modeling framework and offer users improved flexibility to build and manage data models and metric definitions in the semantic layer using code-based modeling frameworks. 

With AtScale Modeling Language, analytics engineers gain enhanced flexibility by allowing them to design dimensional models that represent views of raw data intended to be used for business intelligence and data science.

According to the company, AtScale models include table joins, dimensional hierarchies, metrics definitions, and metadata geared at supporting user interaction from analytics tools.

Next, dbt Metrics Serving offers users an alternative to open-source modeling. This capability allows for the reading of dbt Metrics definitions straight from Git project files. Customers can also establish connections to dbt Models on cloud data platforms, and provide dbt Metrics to AtScale-supported analytics tools. 

This enables analytics engineers to work in an open-source modeling environment while utilizing AtScale’s analytics layer integration, push-down query execution, and automated aggregate orchestration. 

Lastly, the company stated that the addition of Python-based Metrics Engineering to AtScale’s AI-Link offers improved management capabilities for metrics stores. With this, organizations can create, read, update, and delete multiple AtScale objects using a Python API. 

This helps analytics engineers and data scientists manage large metric stores while avoiding the need for manual updates. 

“Analytics engineers and other code-first data modelers need the flexibility of a markup language and automation scripts to build and maintain the sophisticated data models underlying a robust semantic layer,” said Dave Mariani, founder and CTO of AtScale. “AtScale’s modeling language is built on best practices of dimensional analytics and seamlessly integrates with our metrics serving engine, ensuring optimal performance and cost efficiency of analytics queries, while maintaining tight integration with analytics layer tools.”

The post AtScale unveils new capabilities to support code-first data modelers appeared first on SD Times.