Open new doors for legacy software
When it comes to software, industry and vendor changes can leave project managers feeling painted into a corner. Soft-X-Form™ opens new doors by converting existing software to new platforms and languages without rewriting, at very significant savings compared to a full rewrite.
Modeling Software Replacement Costs
The cost to develop software is the sum of the costs to provide individual features.
When it comes to replacement software, everyone wishes that replacement software could be rewritten for a fraction of the original costs. But the fact is that rewriting from scratch is just as costly as new development, because it IS new development. Increased productivity due to developer experience saves just a few percent. The problem is that except for a few special cases[*], the feature set to develop is the same, so the rewrite cost is high.
Using Soft-X-Form™ radically changes this. Using compilers and translation tools, it extracts the functional meaning and purpose (i.e. the features) from existing software source code. It then uses code generators to re-implement that meaning on new languages and platforms.
Using Soft-X-Form™ achieves significant savings compared to a "from scratch" rewrite:
| Software Activity | Rewrite | Soft-X-Form™ |
|---|---|---|
| 1. Request (e.g. requirements) | A complete written inventory and description of features (used for development and testing) and an architecture plan (for the new platform/language) are still required. Unfortunately, most projects, especially those developed with "agile" methods, do not have complete and up-to-date requirements documents that can be used for a rewrite. Typical project portion: 25%. | The "specification of requirements" and business logic architecture come from the existing implementation. Advantages: Avoids re-writing requirements and eliminates the problems caused by using using stale, out-dated requirements documents. |
| 2. Development | Using experienced staff should result in high productivity, but the development size is the same since it still starts from scratch.
In addition, it is often the case that new software features are required (e.g. to support the "automatic transition" of data to the new system from the old.) Typical project portion: 50%. | By automatically translating 95-99% of the source, developers are left with a small number of gaps to bridge manually. (This work can be reduced even further with pre-conversion preparation and custom bridges.)
After conversion, you will be able to maintain the software as you did before. (There is no obfuscation. Comments and structure are preserved.) |
| 3. Test | Testing will require as much effort as the original. The features implemented are the same. Defects still must be located by exhaustive testing in order to be removed.
Typical project portion: 20%. | Testing is required, but is limited to testing the parts touched by developers and to ensure the translation processor itself is correct.
The correctness of the translation processor is validated by sampling, not exhaustive testing. |
| 4. Deployment (e.g. publishing) and user adoption (e.g. training.) | User training (including online documentation, training materials, etc.) must also be rewritten unless the rewrite is a full "form-and-function" replacement.
Typical project portion: 5% (Post-deployment, training, materials, not included.) | Because this is a high-fidelity conversion, it is a full form and function replacment.
User training and materials may need to be updated, but not rewritten. |
It is important to note that software effort models may not take into account the post-deployment costs, which may be significant for both the organization and the down-stream customers.
Back to top | Next: Avoid disruption as platforms evolve
* - Two examples of special case reductions: There is no need to create replacements for software features that were abandoned along the way. Also, utilizing a pre-developed application stack (e.g. SaaS or special-purpose language) can provide many features at a fixed or leased cost, rather than developed cost. This may save 30-50% of the size to be rewritten, but impact the long-term licensing and operational costs.