Question about Oxygen Author performance

[akiraaa]

Hi there,

We are using a DITA topic for conref and conditional output, and now this file has around 4,000 lines (200 KB). In the future this file may reach tens of thousands of lines, we are wondering how Oxygen Author performs for such large files. Will it be tenable to open and edit such a file using Oxygen?

Thanks!

[Radu]

Hi,

We have clients editing DITA topics in the range of 5-10 megabytes in Oxygen, of course the loading of the topic takes more time as the topic increases in size. In general you should consider splitting your reusable content in multiple topics, maybe based on certain criteria, a topic for reusable notes, one for reusable tables for example. Oxygen's "DITA Reusable Components" view has a "Components" tab in which it should present all such reusable elements no matter if they are defined in separate topics or not.

Regards,
Radu

[AngelaSimonelli]

Hello,
I have a related question regarding Oxygen Author performance. Many of our users frequently experience significant slowness and lags while modifying DITA topics that range in size and complexity. Some users experience lags while they're simply typing and scrolling as well as when they're reorganizing content within a topic, etc.

Judging by the Oxygen requirements discussed in another post (post19979.html#p19979) I think our machines are more than adequate to run Oxygen and many simultaneous tasks (https://ark.intel.com/content/www/us/en ... 0-ghz.html).

I am curious about what you would say constitutes a "very large file", with respect to the guidance not to give Oxygen more than half of the system memory (e.g. -Xmx2048m for 4GB of memory). As Adrian Buza wrote, "You can change the amount of memory available to Oxygen by adjusting the value of the -Xmx argument in the .vmoptions file... But only do this if you are working with very large files."

Our setting is Xmx3229m for most users and we did increase computer memory and Oxygen memory allocation to Xmx4000m for at least one user who experiences daily Oxygen lags and slowness.

Among DITA topics in which users have reported experiencing lags between keystrokes while typing, scrolling, and taking other simple actions, file sizes were:
[*]425 KB (and file contained 20 -- 50 tracked changes)
[*]683 KB (and file contained 50+ tracked changes)
[*]15 KB (and the file contained ~10 tracked changes... but 40+ links to other DITA topics and maps

Does any of this data help to explain the performance lags? What else would you recommend we investigate to help solve the problem of our users experiencing slowness and unresponsiveness while working on DITA topics in Oxygen?

Thanks very much for any assistance,
Angela

[Radu]

Hi Angela,

What version of Oxygen are you using?
Do your computers have a hard drive or a SSD (solid state drive) installed?
Your DITA files are not that large, also if there are lags they are usually when opening the file, not when editing it.
For example the last lag I profiled was a 15 MB XML document which took about 3 minutes to open, but editing worked well after that, also that file contained thousands of mathematical equations, so it was something quite huge.
Are the XML documents you are opening on the local disk or on some kind of content management system?
Do you have any add-ons installed in Oxygen? Because that could explain the lags, add-ons/plugins can be notified whenever users modify the document and we do not control what those add-ons are processing while the end user is modifying the document?
Usually lags while editing can be explained either by some intense processing which is going on or by a network operation, if for example there are resources which Oxygen needs to read from a remote server. For example if you reference images from an HTTP server,but Oxygen has various caches in place even for this...
If you can reproduce such a problem without any add-ons installed (Oxygen Preferences->"Plugins" page) and with files placed on the local disk, maybe you could share with us a sample file and some steps to reproduce the problem. In the Oxygen main menu Help->"Support Tools" there is a tool to randomize the text content of XML documents, as we are mostly interested in the XML structure, not the original text content itself.

Regards,
Radu