| can only read his MAT file in a specific Matlab version | uses MySQL-Matlab interface to access his data from all Matlab versions |
| cannot access his data easily from C, Perl, Python, R, command line | can access his data from most languages which have a MySQL interface |
| is limited by Matlab memory in how much data he can query | is only limited by the size of his hard drive |
| looking for specific data points involves a slow linear search of all data | indexed columns allow very fast searches by only looking at likely matches |
| complicated queries over many columns and rows in a matrix are hard to code | use standard syntax designed for multi-column and multi-table queries |
| everything is a double precision number or a string | freedom to choose the appropriate data type from many possibilities |
| update of a single number requires loading and saving the entire mat file | updates are optimized to only touch the required locations |
| needs to translate mat files and learn a new interface when a tool other than Matlab is needed | uses the same SQL syntax from multiple platforms |
| needs to take his mat files on the road | can connect to his MySQL server from anywhere |
| has to pay big bucks | MySQL is free |
| uses way too much memory and disk space | optimized better than you could ever duplicate yourself |
| only one application can access/modify the data at once | acquisition, automated and manual analysis, web-based systems can all access and modify the data simultaneously without corruption |
| |