Frequently Asked Questions
Are your tools as good as you say?
1. Do you really achieve 85-95% line-level accuracy in legacy code refactoring?
Originally conceived in 1985, our conversion tools have been refined on millions of lines of client code, resulting in an average line-level accuracy rate of 85%-95%.
2. How do you know that the translated code is functionally-equivalent?
Our automated translation tools use a proprietary translation process to ensure functional equivalence. The legacy code is first converted into an intermediate tertiary language. This tertiary language, which is common to our entire family of translators, preserves functional equivalence by maintaining a 1:1 correspondence in concepts and functions wherever possible. In cases where exact correspondence is contextual, our proprietary algorithms, refined over millions of lines of code, come into play, resulting in unmatched accuracy and minimized long-tail risks.
3. How long does it typically take for your clients to generate modern code that compiles?
Depending on the size of the project, a manual conversion that would ordinarily require months or years can be done in just a few hours, days or weeks. The turn-key translation process requires minimal manual intervention. The output source code is ready to be tested, validated and debugged with the target compiler of your choice.
4. Do you offer a demo or proof of concept to prove that your tools would work for me?
We offer a free demo or proof of concept. All we need is sample file of 1,000 lines of code or less (including Include files) in zip format. If the sample file is larger than 1000 lines, we will provide you with a translation of the first 1000 lines only. If your code is sensitive, we support PGP encryption and have even accepted samples via mailed CD/USB. We will sign an NDA upon request.
5. Do you offer a Warranty on the quality of your code translation?
For Warranty information, please see our Standard Terms and Conditions at: https://www.recodex.co/terms
6. Does the RecodeX tool or physical licensing key present any Cybersecurity risk?
No, the tool and dongle are proprietary and manufactured and supported in the U.S.
7. Why can’t I just use Generate AI to modernize my legacy code?
AI-based solutions achieve only 40-60% line-level accuracy on legacy languages like COBOL or Assembly. Our automated tools refactor legacy code in days/weeks, not months/years. They are not just a “coding assistant” and achieve 85-95% accuracy on the first try. Our algorithms are deterministic (not a black box), ensuring functional equivalence and no missed requirements or hallucinations. Also, putting your proprietary code into public AI models puts your data and IP at risk. When using our automated tools, all steps in the translation process are performed by your own engineers on-premises in a non-prod environment, so you retain full control over your code at all times.
Can RecodeX help me?
8. How much does RecodeX charge?
Automated translation costs 5-10% of manual application refactoring. Our software conversion tools are cost effective, being priced 25-50% less than transpiler competitors. Our charges are based on the size of your conversion project, i.e. the number of lines of legacy code to be translated. For an immediate, free-obligation quote, please use the contact form on our web site to tell us about your legacy software modernization project.
9. What if my legacy application’s source language is not on your list of supported languages?
If you cannot find your legacy source language or exact dialect listed on our web site, we may be able to modify an existing tool to suit your needs (in fact that is how many of our tools were originally developed). We will select the closest translator to your legacy program requirements and use it as a base line. Then, we will modify the front-end of that translator tool to accept your legacy language or dialect. Give us a call or send us a small sample of your legacy code so we can estimate the modification costs involved in creating a translator that will modernize your legacy source code.
10. What if I want to refactor to a target language that is not on your list of supported languages?
We will select the translator for your source code language and use it as a base line. Then, we will modify the back-end of the translator to output your target language. Give us a call or send us a small sample of your code so we can estimate the modification costs involved in creating a translator that will output the target language you want.
11. Do I need to be moving to a specific platform (i.e., AWS, Unix) to use your tool?
No, you can translate to any of the target languages that we support and then choose the appropriate compiler or target platform. Our tools work no matter what infrastructure you choose – mainframe, cloud, public, or private. Our translation tools use a proprietary Windows interface: You select the source language (including dialect); you select the file(s) to translate (including any Include/Copy files); and then you view the translated output file(s). Our resources are focused on always improving the performance of the translators and not on the operating environment of the translators. If you have a very large translation project (350k lines and up), a new operating environment may be justified. Please contact us for a no-obligation quotation for modifying the translator that you need to work in the operating environment that you desire.
What do I need to get started?
12. What hardware and software are required in order to run RecodeX’s tools?
The translator is a Windows console program. The translator will run on any computer with Windows XP and higher with a USB port.
13. How do I use the physical licensing key?
The physical licensing key simply gets plugged into the host computer’s USB port.
14. How many engineers do I need to dedicate to working with RecodeX and for how long?
Because the translation process is automated and streamlined, you can do a large conversion project with just 1 or 2 developers in hours or days, instead of months or years.
15. What level of manual intervention is required?
The turn-key translation process requires minimal manual intervention.
16. Do the developers need to understand COBOL (or other source language)?
The automated translation process requires little to no knowledge of the source language, i.e. COBOL.
17. How can I improve the accuracy of the code generated by the translator?
For most applications, the standard translator should be sufficient. However, if the resulting code is not satisfactory, your developer should provide us with a list of issues in the translation output, and if the error(s) is on our end, we will modify the tool to improve accuracy and provide you with an update. Otherwise, for an additional charge, we will customize the translator to your style of programming in order to enhance and optimize the quality of the outputted code and to reduce the amount of work needed in the verification and validation process. If required, remote or on-site consulting can also be arranged.
18. What compiler do I use? How do I validate the translated code?
Our translators are source to source compilers that accept software code in many dialects and convert it to C, C++, C#, JAVA, J#, or VB. You can use the target compiler of your choice.
The output source code is ready to be tested, validated and debugged with the target compiler of your choice. The comments are moved transparently from the source file to the target file. Any built-in functions for math, string manipulation and I/O in COBOL (or other source language) are converted to calls to external procedures.
19. Since it is difficult to onboard new vendors at my company, is RecodeX willing to work as a subcontractor?
Yes, in assisting with translation projects, we have worked as a subcontractor to numerous companies and government contractors.
Can you answer my technical questions?
20. What is the largest size legacy application your RecodeX tools can modernize?
Our tools can easily convert and deliver more than 2.000,000 lines in a week.
21. What is the largest individual file size that your tools can refactor?
We have successfully converted files as large as 250,000 lines to 500,000+ lines without any problems with any limitations depending on the memory in your system. We have a number of different translation options that will conform to the size of your source file.
22. Which Include files should I use with the translator?
When you use the translator, all of the Include files that you use must be in the source language.
23. What if I don’t have an Include file?
You can still translate a file when you do not have the Include file. The translator will give you a warning for every missing Include file and will implicitly declare any variable that is not pre-declared or that is missing. However, in order to optimize your translation, it is preferable to have all of the Include files on hand.
24. Do your tools also migrate legacy databases?
If your code has one or more embedded statements from legacy databases in the source language, the translator gives you several options:
- Transfer the statements to the target as they are and use an SQL pre-processor for the target language.
- Pass the statements to the target language as comments and modify them later on for the new data base connectivity.
- Call an external SQL, DB2,ASM, CICS translator to translate the embedded statement.
25. How do you handle errors in the source code?
The translator flags syntax errors in a listing file with levels for Include or Copy files. The translation may be done in an editor environment with error message capability using "Brief" or any editor that is supported. The translator also has a by-pass capability that allows translation even with syntax errors in the input file.
26. How do you handle in-line comments in the source code?
The translator can convert source files with comments passed through and logically placed.
