Azur Systems vs. Dryfta

I would highly recommend Azure machine learning design for those with less access to high-end computing infrastructure, as using Azure saves a lot of time, money, and effort by providing a hustle-free platform that is easy to use and train your employees on. On the other hand, if you are looking for complete control of the machine learning model you create and would like to add detailed functionalities and try different algorithms, then Azure is less suitable here as it’s very high level.

Incentivized

I can't provide a scenario where Dryfta could be well suited. I guess it's a matter of delivering what has been promised on time, and without having to invest thousands of hours in extra work, as it was in our case. If it worked properly it could be a good tool for any conference.

• Leverage machine learning as a service.
• Accelerate machine learning with best-breed algorithms.
• Support remote working with cloud-based services.
• Creation of compliant and secure machine learning applications.

Incentivized

• It has a lot of functionalities (website and administration) all included in the same platform. If it worked properly, it allows you to organise a conference paper-free.
• Help desk responds quite quickly
• They consider making improvements when asked for.
• It’s not too expensive.

• Limited storage for those using the free version.
• Limited number of machine learning algorithms.
• Because of its high-level design, it doesn’t give the developers much control as you don’t write much code in to do what you exactly want.

Incentivized

• Multilingual issues. Although it is advertised as multilingual, it didn’t really work as such. The many issues that popped up throughout the conference preparation were fixed little by little at a cost in time, from help desk emails and to struggles with the unfriendly UX.
• Admin pages reloaded every time you clicked on a button (their developers seem to ignore Ajax technologies). It was time-consuming and required constant page searches.
• Inflexibility in many of the supposed functionalities it offers.
• Certificates were not modifiable nor custom when we had to send them (it was solved months after the conference finished when we were surprisingly contacted by the help desk).
• Problems with the size of images to be displayed on the site, very small fonts and limited options to display content. We had to hire a professional developer in order to get a graphically consistent and presentable website.
• Very poor mobile version. Too big margins, unreadable text, endless text blocks and lists, distorted pictures, etc.
• Issues with the ordering of the authors’ names for different proposals (authorship being so important in research).
• Fixed, inflexible fields in the contact sheets, speakers info, and so on.
• Special character issues (due to Latin characters and other types used in linguistic research).
• Not being able to include links in the HTML editor due to Dryfta's inadvertent decisions to block them.
• Only one Superadmin user allowed to access the full functionalities of the platform, so we had to share it (consequently not knowing who did each action).
• Problems with the generation of reports and the high complexity of their interface.
• Some issues on the mandatory anonymity. The double-blind review process not fully respected due to unclear user info and options, with other issues coming up on the go such as unwanted info in automated notifications and messages in the Welcome dashboard.
• Not being able to use the other payment methods on the platform because they were incompatible with the conference country.
• Missing information and time wasted when creating events for sessions with info that already existed in the server that randomly failed to be selected. These issues were reported even with video proofs (help desk didn’t believe us), and were never solved. We had to repeat the same processes again and again, never knowing what was going on

It was quick, that's all we can say. Quite a few times they sorted out the problems and issues. But, sadly, sometimes their answers were useless and irritating (not addressing the problem or simply ignoring it, "passing the buck"). In some cases, they pretended the problem didn't exist and we had to send them videos as proof. No response to that.

It’s fierce competition, but Azure as a Microsoft product is more stable and performs better than Amazon AWS since all of our machines are built on Microsoft Operating systems. Azure provides a pay-per-minute service, unlike AWS, which provides a pay-per-hour service. Azure allows direct linking to exist cloud storage that we already use at our institution, such as OneDrive. Azure has a friendly and easy-to-use interface that staff and students are already used to working with other Microsoft products.

Incentivized

• Reduce energy consumption caused by GPUs.
• Saves on recycling and transporting costs and maintenance caused by buying high-end equipment.
• Improve productivity as building products using Azure is easier than building everything up from scratch (e.g., machine learning and AI applications).

Incentivized

• The overall performance was okay in the end, but it was due to our team's commitment and effort. Without that extra work, the results would have been awful.