The post Beyond New Features: The Unsung Work Behind Long-Term Support appeared first on Collabora Online and Collabora Office.

Save As?

TL;DR
When you click “Save As” in your favourite word processor, you probably don’t think too much about the file format. Yet file formats like .doc, .docx, and .odt shape how we create and share documents—but they can also create headaches. Here’s a quick guide to their history, differences, and why Collabora Online makes working across them easier.

Related content:

Podcast

What is a Document?

Listen to the podcast now

From Closed and Binary, To Open and Extensible

The earliest digital files were of course just plain text – a simple sequence of characters, and no formatting. Writers used to all the flexibility of pen and paper however quickly wanted more: bold headings, paragraph spacing, indents, embedded images, at this point the list is almost endless. This led to the evolution of file formats storing much more than the mere ‘text’ of a file. Binary formats such as WordStar “Document” files (introduced 1978 – interestingly WordStar files had no official file extension, files were either “Documents” with formatting, or “Nondocuments” pure ASCII text files), or Microsoft’s .doc (1983) took over the word-processing space. Compact and efficient for older systems, binary formats also locked users into specific software, making it difficult to share documents across platforms.

Above: What’s going on under the hood?! The same document saved 3 ways

By the 1990s, Microsoft Word dominated the market, cementing .doc as the “default” document format worldwide.

The next big shift came with XML-based formats – more verbose, but human-readable and compressible. The open-source community were the first to develop the Open Document Format (ODF) in 2005, designed to prioritise openness, readability, and extensibility. Microsoft soon followed suit with their own OOXML files such as .docx in 2007, while ODF continued as the ISO-standardised alternative.

Whereas binary .doc files were proprietary containers that embedded program-specific data structures directly, ODF and .docx files are zip archives full of multiple XML and media files, making them easier to parse, extend, and standardise than their opaque binary predecessors. You can think of .doc as a box not only locked, but also containing pages of Egyptian hieroglyphs – only Microsoft held the key and translation documentation, and they were unwilling to share. ODF files (and largely .docx*) on the other hand are transparent zip files written in your language. You can open them, inspect the contents, and tweak them as much as you like.

*Reality check on OOXML

There’s no smoking gun that .docx was engineered to be opaque. However, OOXML’s sheer breadth, the long-standing default use of the “Transitional” profile, and historical standardisation controversies have raised questions and interoperability costs for alternative suites putting a question mark over some of the claims to being an ‘open’ format.

In practice, this means that while .docx is technically an open standard, third-party developers often face hurdles when trying to implement full compatibility.

Don’t change formats while editing?

For open standards and long-term access, ODF (.odt, .ods, .odp) is ideal. As an interesting case study, and illustrating the benefits, many governments now mandate the use of open file formats as the “standard for use across government”.

There are many reasons for this, but they essentially boil down to making sure you have full control over your documents – and avoiding the technical and financial dangers associated with proprietary technology. Essentially if you are relying entirely on one proprietary company for your file format standards, document editing software, and perhaps even cloud storage, it could be argued you’re asking for vendor lock-in problems, with future migration or even simple access becoming costly.

Legacy proprietary formats (such as .doc or .rtf) which are no longer actively developed should be avoided. In the future, if not already, they will be increasingly difficult to work with (have you tried to preserve formatting while opening Microsoft Works .wps or .xlr file in a modern document editor recently?).

The elephant in the room is what to do with .docx, .xlsx, and .pptx files. As proprietary formats, they carry the same long-term risks as .doc or .wps – future deprecation, compatibility gaps, or costly migration. Yet in a world dominated by Microsoft, ignoring them would be impractical. The safest approach today is to continue using these modern proprietary formats (without conversion) with existing files, but to prefer open formats (ODF) for creating new content. This strategy minimises the danger of future lock-in and preserves your freedom to move documents between systems without technical or financial obstacles.

What about Document Editors?

This all means you need a document editor that is as comfortable working with .docx as it is .odt., .pptx or .odf. The good news is Collabora Online natively opens, edits, and saves all major formats, both Microsoft’s and ODF, so you can edit and collaborate freely without conversion headaches.

• ODF formats (.odt, .ods, .odp) are ISO-standardised, open, and widely supported – making them the safest bet for long-term storage and interoperability.
• If you’re exchanging documents predominantly with Microsoft Office users, then .docx, .xlsx, or .pptx may be more practical. Those formats are the default in Microsoft’s ecosystem, and choosing them avoids confusion when sending files back and forth.*
• Legacy binary formats (.doc, .xls, .ppt, .rtf) should be avoided for new work — they’re deprecated and increasingly unreliable.

The takeaway? File formats will keep evolving, but you don’t have to get stuck in the past or tied to a single vendor.

Collabora Online does a better job of handling Microsoft formats than Microsoft does of handling ODF. That means if you stick to ODF in Collabora Online, your documents will remain open, future-proof, and feature-complete – but if you need to share in Microsoft formats, we’re ok with that too, and provide high-quality export options that preserves your work.

The Best of Both Worlds

Get the best of both worlds with Collabora Online!

Try Collabora Online today and experience smooth, format-agnostic collaboration.

Free Demo

The post A Brief History of File Formats: DOC vs DOCX vs ODF… appeared first on Collabora Online and Collabora Office.

Fixing a critical loadurl performance issue under Android

• July 8, 2025
• Mobile, Support
• No Comments

TL;DR

Stalls in document loading were traced back to issues with the long-string handling abilities of Android’s Webview loadUrl method. Implementing an alternative websocket-based approach eliminated the problem.

Introduction

Everyone wants snappy scrolling and quick loading documents, so a design issue in the Android Webview loadurl method appeared to make the Collabora Online Android app fail to load documents was of critical concern, as it seriously impacted the end-user experience.

A user-reported GitHub issue seemed to indicate that larger documents were failing to load or only load in part. The initial investigation into the problem highlighted that the root of the issue wasn’t that the document failed to load, more it was rendering impossibly slowly, giving the appearance of a load failure. Once the issue was confirmed, it was tagged as a high priority and assigned to our Android developer Skyler to fix.

Behind the scenes

Collabora Online is available for both iOS and Android devices. To enable the mobile app to function offline or with interrupted or poor network connections it utilises its own local coolwsd – the server that renders and streams image tiles of your documents to the client so they can be viewed and edited. Coolwsd renders the documents being viewed into 256×256 raster tiles, with the original reported implementation these are base64 encoded – to safely transport the UTF8-encoded strings – and loaded by the browser client. This should be a smooth uninterrupted process, which it was not.

Our assigned developer Skyler implemented performance profiling to try and narrow down and locate the issue. The first move was to run testing on iOS to see if it was suffering the same performance issue, it wasn’t, so this appeared to be an Android-specific issue.

Profiling the issue

The first step in assessing the issue was to devise suitable profiling and debugging. The initial flame graph for the system seemed to indicate that either the system was sending way too many tiles or the tiles that were being sent were taking an age to be processed.

Analysing the browser logs revealed 105,707 tile messages, so there were a large number of tiles being processed but none of them were actually making it to the user’s screen. At this point it made sense to see if it was possible to optimise the situation either through better prioritising or removing duplicate requests, then see if that alleviated the issue.

Further log analysis showed that a whopping 91% of tiles were duplicates. Implementing a patch to remove duplicates still led to an enormous number of invalid tile requests. After testing under the Core and the Desktop versions, Skyler logged to see if tiles were being sent unrequested and they were, alongside a large amount of duplicate tiles.

After investigating the duplicate tile issue, Skyler developed a prototype patch to help reduce the number of duplicate tiles resulting in a reduction in lag, but the rending speed still remained slower than it should.

LoadUrl was the culprit all along!

The actual culprit for the performance hit turned out to be Android’s Webview loadUrl method, which doesn’t like handling very large strings. The encoded tiles have to be passed to the browser in the form of a very large string as part of a Javascript function through a fake websocket. An initial solution was to send anything over 1,024 bytes through a HTTP request with a short message still going through Javascript to the Webview. This led to the final patch set that created an Android MobileSocket, based on the existing ProxySocket – implemented to work around PHP’s timeout limits and situations involving strict firewalls that don’t allow websocket connections.

The most striking way to experience the difference in performance is to watch the before and after videos, which are night and day in performance terms. Delving into the system tasks the main culprit was the WebViewChromium.loadUrl task, this was in effect stuck trying to load tiles until the user gave up on it. The final fix still uses this task, but only for message polling all the tile data goes through the MobileSocket.

The end result is vastly faster document loading when viewed against the earlier releases. You can find the 25.04.2 Android release of Collabora Office, which includes this work, on the Google Play Store.

Try it yourself, get the apps!

Download Collabora Office for iOS and Android:

iOS Apple App Store

Android Google Play

Download the Android APK

Learn more

Collabora Online Development Edition

Head over to the docs or start by following our step-by-step instructions and build CODE from scratch.

Get involved

Working with other hackers on improving Collabora Online is fun and very rewarding! Find out how you can get involved.

View easy hacks

Subscribe to our newsletter:

Follow us on social media:

Github Linkedin Mastodon Youtube Fontastic-icon-bluesky-logo Twitter Facebook Instagram

Developer Blog

Export vector formats in background image to pdf

Read More »

Collabora Office

Interop EXPLAINED

Read More »

Developer Blog

Ensure scrolling has smooth animation

Read More »

The post Fixing a critical loadurl performance issue under Android appeared first on Collabora Online and Collabora Office.

At Collabora Productivity, we like to be productive by name and by nature. One recent feature fix for Danish insurance company – balticfinance, has shown this to be the case. Ten working days after being requested, the updated pivot table functionality was delivered.

Pivot tables are an extremely powerful feature, to reorganise (or pivot) data from one layout into another. In the image below for example, the data from columns B-D has been reorganised by ‘Category’ and the ‘Job#’ into columns F-H, with the Value of each Job automatically summed together at the bottom. As an additional convenience, each category could be collapsed/expanded by clicking on the – or + symbols to view the values itemised or summed for the category.

As the cells within a pivot table have been generated by the spreadsheet, the contents cannot be directly edited. They instead mirror the value from the cells they are referencing, or automatically sum a category/total. This also makes cell formatting within pivot tables more complex than other cells. While Collabora Online has had strong interoperability with pivot tables created in Microsoft Excel for some time, previously we did not display cell formatting (such as background colours) for pivot tables created in Excel. balticfinance came to us with this exact request.

After receiving the support request on Friday 26^th January, Collabora engineers began work on the update on Monday the 29^th, with the fix delivered for testing on the 9^th February – two weeks from the initial request. The fix now means pivot table cells edited in Excel and opened in Collabora Online both retain their formatting, and survive the round trip export back to Excel in Microsoft’s OOXML format.

As well as being a great example of the deep interoperability offered by Collabora Online, we feel this story well illustrates the benefits of working with a responsive company such as Collabora Productivity for your office solution. Unlike some faceless companies, we are open and want to work with you.

If you have a special use case – get in touch! Let us help you free your documents. Learn more about our Consultancy or contact us at hello@collaboraoffice.com.

The post New Features in 2 Weeks? A Success Story of Productive Pivoting Spreadsheets appeared first on Collabora Online and Collabora Office.

The need for new types of certification reflects the growth and maturation of businesses supporting LibreOffice, as well as the leadership of the Document Foundation and its readiness to meet that need. Over the last two years The Document Foundation have granted certification to 45 software development engineers. Graduates include employees of leading Open Source multi-nationals Red Hat, SUSE, and Ericsson.

As a non-profit organisation with charitable status, The Foundation provides a degree independence and authority which corporate bodies can never achieve. With its three-tier management structure, TDF is well suited to non-partisan setting of standards and assessment of certificate applicants. Yet because many leading LibreOffice service providers are also board members, the foundation is uniquely positioned to assay best practice and learning resources, and foster cooperation between companies which may otherwise be marketplace competitors.

“In fact, LibreOffice Certification is the first of this kind to be managed by a community based Free Software project, as all other certifications in the open source environment are managed by a company” — Italo Vignoli, Certification Committee Chairman

The peer-to-peer certification process covers a broad range of themes from “certification theory” to “growth potential”. Some subjects are special to the application’s Open Source roots, including “basic knowledge and understanding of Free Software communities”, and “Free Software licenses”. Certification both incentivises professional learning, and provides a public directory of qualified experts for those seeking services related to LibreOffice.

Four of the six companies with newly qualified staff are Collabora partners, and 13 of 45 engineers already certified are members of our team. With more certified staff than any other company, we’re delighted to be part of the training and development of LibreOffice experts, and look forward to promoting certification to our international network of educators and integrators.

The post Why LibreOffice Certification Matters appeared first on Collabora Online and Collabora Office.