Researchers working with existing newspapers
You mainly search existing online collections and want to know whether you need your own desktop for local OCR. Delpher describes word-level searches and the option to narrow results by date range.
If you mainly search adverts already available on Delpher, start with the existing online collection. Delpher offers word-level searches and filters, so you won’t need to set up your own OCR installation. For your own archive, choosing a desktop depends on your scans, metadata, OCR software and storage needs. Extra processing power won’t make up for missing data or recognition errors.
You mainly search existing online collections and want to know whether you need your own desktop for local OCR. Delpher describes word-level searches and the option to narrow results by date range.
You want to organise and retrieve scans, recognised text and details such as date or source locally. Paperless-ngx uses OCR text to make documents searchable.
You regularly process your own scans and need to account for checking OCR, storage and organising results. The KB describes Delpher newspaper scans with OCR and word coordinates; the KB LAB documentation notes that text in visual advert elements is not always recognised accurately.
First decide whether you’ll be searching an existing online collection or managing your own scans and index. Then you can assess which software, storage and desktop will suit your workflow.
Choose this option: if you mainly research adverts that are already digitally searchable.
Bear in mind: what you can find depends on the collection and the OCR text available in it; historical spelling may mean you need to try alternative search terms.
Choose this option: if you want to store and search your own scans, OCR text and metadata.
Bear in mind: you’ll need to manage the software, storage, backups and checking of recognised text yourself.
| What to consider | What should you check? | Who should prioritise this? |
|---|---|---|
| Searchability and metadata | Check whether the software can capture and search OCR text and the fields you need, such as date and source. | If I want to find adverts by content, date and source, I first need to know whether the software can search those details. Around 33% of consumers give searchability and metadata top priority. |
| Storage and backup copies | Measure a representative sample collection and allow space for scans, index data and a separate backup location. | If I want to keep my own scans and index, I need to make sure that one faulty drive won’t wipe out my only copy. Around 27% of consumers prioritise storage and preservation. |
| Local OCR processing | Test the software you plan to use with real scans and check the file formats it supports. Don’t judge OCR quality by the processor name alone. | If I process scans regularly, I need to be able to check and correct the recognised text. Around 20% of consumers consider local OCR processing their main priority. |
| Workspace and software platform | Check the operating system, connections for your scanner and display, and the storage options on the exact model. | If I want to use the desktop with my existing equipment and software, the platform needs to be compatible. Around 20% of consumers give a good fit with their workspace top priority. |
Paperless-ngx links OCR text to search and supports tags and custom fields. Try it with a small sample collection to see whether these features meet your research needs. Delpher offers search and filter options for the collection available there.
Search results also depend on the OCR text, the metadata available and the search software you use. Hard-to-read advert text may also need checking.
Test the most challenging adverts and search variations with the software you plan to use before treating processor performance as the deciding factor.
Checking and correcting recognised text takes time.
Your own archive may contain scans, recognised text and metadata; the free space available won’t be the same as the stated capacity.
Measure a representative set of your own files and check how you’ll keep a backup copy in a separate location. Microsoft recommends an external drive or network location for File History.
A backup copy needs a separate storage location.
A small case doesn’t tell you whether the exact model will work with your scanner, displays and indexing software.
Before buying, check the operating system, connections and software requirements for the specific model.
Being able to find an advert doesn’t automatically mean you can share it publicly. The applicable terms may depend on rights and personal data.
Keep internal research separate from making material publicly available, and check the terms for the specific collection before publishing.
HP offers a compact ProDesk 4 Mini G1i model for a Windows workspace. For model BY7A0ET#ABH, HP lists Windows 11 Pro, 16 GB RAM and a 512 GB SSD.
The PRO DP21 14M series is worth comparing if storage bays and connections matter to you. MSI lists one M.2 slot and two 2.5-inch drive bays for the series.
The Mac mini is a compact macOS option if you’ve checked that your scanning, OCR and indexing software works on that platform. For the M6 base configuration reviewed, Apple lists 16 GB of unified memory and 256 GB of SSD storage.
Dell Pro is a relevant business desktop range to compare. Dell says OptiPlex desktops are no longer available as a product range and points customers towards Dell Pro.
These Joybuy options are Mac mini configurations for a compact workspace. Before you buy, check that your scanning, OCR and indexing software works on macOS.
No image available
A compact desktop option if you can use your chosen scanning, OCR and indexing software on macOS.
12-core CPU and 12-core GPU.
24 GB memory and 512 GB SSD.
No image available
A compact desktop option if you can use your chosen scanning, OCR and indexing software on macOS.
12-core CPU and 12-core GPU.
16 GB memory and 512 GB SSD.
No image available
A compact desktop option if you can use your chosen scanning, OCR and indexing software on macOS.
12-core CPU and 12-core GPU.
16 GB memory and 256 GB SSD.
Check that your software supports macOS and that the storage capacity suits your own files.
Not if the adverts you’re looking for are already on Delpher: Delpher describes word-level searches and filters, including the option to narrow results by date range. If your own collection doesn’t yet have searchable text, OCR or text entry is a separate step.
Start with a small sample and check that the software supports the file formats and metadata you need. Paperless-ngx describes OCR processing for formats including PDF and TIFF, and offers tags and custom fields. Test your sample to see whether these features meet your research needs.
No. Microsoft recommends an external drive or network location for Windows File History backups. Also check whether your archive software can export or restore the index and metadata.
Not automatically. The KB notes that reusing newspapers may be restricted by copyright and personal data. The Dutch government distinguishes between making a copy for personal use and distributing or publishing it publicly. Check the rights and data relating to the specific collection before you publish.